SEARCH

SEARCH BY CITATION

Keywords:

  • Biopsy complications;
  • glomerulosclerosis;
  • hepatorenal syndrome;
  • kidney allocation;
  • kidney function;
  • liver transplantation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Kidney Allocation
  6. Results
  7. Discussion
  8. Disclosure
  9. References

The feasibility, value and risk of percutaneous renal biopsy (PRB) in liver transplant candidates with renal failure are unknown. PRB was performed on 44 liver transplant candidates with renal failure of undetermined etiology and glomerular filtration rate (GFR) <40 mL/min/1.73 m2 (n = 37) or on renal replacement therapy (RRT) (n = 7). Patients with ≥30% interstitial fibrosis (IF), ≥40% global glomerulosclerosis (gGS) and/or diffuse glomerulonephritis were approved for simultaneous-liver-kidney (SLK) transplantation. Prebiopsy GFR, urinary sodium indices, dependency on RRT and kidney size were comparable between 27 liver-transplant-alone (LTA) and 17 SLK candidates and did not relate to the biopsy diagnosis. The interobserver agreement for the degree of IF or gGS was moderate-to-excellent. After a mean of 78 ± 67 days, 16 and 8 patients received LTA and SLK transplants. All five LTA recipients on RRT recovered kidney function after transplantation and serum creatinine was comparable between LTA and SLK recipients at last follow-up. Biopsy complications developed in 13, of these, five required intervention. PRB is feasible in liver transplant candidates with renal failure and provides reproducible histological information that does not relate to the pretransplant clinical data. Randomized studies are needed to determine if PRB can direct kidney allocation in this challenging group of liver transplant candidates.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Kidney Allocation
  6. Results
  7. Discussion
  8. Disclosure
  9. References

Since the implementation of the model of end-stage liver disease (MELD) system in February 2002, the number of simultaneous-liver-kidney (SLK) transplants has more than tripled from 135 in 2001 to 439 in 2007 with SLK transplantation now accounting for almost 7% of all orthotopic liver transplants in the United States. (1). This trend has been attributed not only to the incorporation of serum creatinine in calculating the MELD score but also to the well-documented negative impact of preliver transplant renal dysfunction on postliver transplant outcome both in the pre- and post-MELD eras and the assumption that SLK transplantation will eliminate the adverse effects of pretransplant renal failure (2–7). However, SLK does not always confer a survival advantage to liver transplant candidates with renal failure. Gonwa et al. analyzed the UNOS data base and showed equal 3 years patients' survival between liver-transplant-alone (LTA) recipients with serum creatinine >2 mg/dL who were not on renal replacement therapy (RRT) at time of transplantation and those with similar degree of renal dysfunction who received SLK transplant (69.8% vs. 69.9%, p = 0.18) (2). Interestingly, this analysis also revealed that 29% of SLK transplants were not on RRT and almost 15% had a creatinine < 2 mg/dL at time of transplantation suggesting that some patients received SLK without a clear survival advantage that raises a number of unanswered questions (2).

A significant number of liver transplant candidates with renal failure recover kidney function even before liver transplantation. Davis and colleagues showed that 6.5% and 15% of patients on RRT at time of listing for SLK and LTA transplants, respectively, were not on RRT at time of transplantation (8). Postliver transplant renal function recovers also in 58–96% of patients with hepatorenal syndrome while patients on RRT for more than 8 weeks duration have a low likelihood of renal functional recovery and benefit from SLK transplantation (9,10). However, kidney allocation to candidates presenting for liver transplant evaluation with renal failure and overlapping features of both hepatorenal syndrome and chronic kidney disease (CKD) constitute a particular challenge due to the difficulty in estimating the extent of irreversible renal damage and the potential of posttransplant renal function recovery in this selected group of patients (8,11,12). In this selected group of patients, it has been suggested that histological evidence of advanced renal damage will portend a low likelihood of postliver transplant renal recovery and will identify those who would benefit from either LTA or SLK transplants (5,6,8,12). Preliminary results from centers with small number of patients and with limited posttransplant information suggest that histological criteria can be used in selecting SLK transplant candidates (13–15). Nevertheless, kidney biopsy has not been widely incorporated in the evaluation of liver transplant candidates with renal dysfunction primarily due to the increased risk of bleeding.

We sought to examine the feasibility, value and risks of percutaneous renal biopsy (PRB) in liver transplant candidates with moderate-to-severe renal failure in whom the potential reversibility of renal failure could not be determined with the available clinical information. We hypothesized that relying on histological information would be useful in selecting patients best suited for either LTA or SLK transplantation.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Kidney Allocation
  6. Results
  7. Discussion
  8. Disclosure
  9. References

Study population

This is a retrospective analysis that included 44 liver transplant candidates with impaired renal function of unknown etiology who underwent a percutaneous native kidney biopsy as part of their liver transplant evaluation between August 2005 and January 2008. Review and analysis of the clinical records from these patients was undertaken with the approval from the Mayo Institutional Review Board. Biopsy indication included; dependence on RRT (n = 7) or iothalamate glomerular filtration rate (GFR) <40 mL/min/1.73 m2 (n = 37). GFR was assessed using the clearance of subcutaneous nonradiolabeled iothalamate during water diuresis (16). Results were corrected for body surface area and expressed as mL/min/1.73 m2. In six patients, iothalamate GFR was not available and was replaced by 24-h urinary creatinine clearance as estimated GFR using different equations is not accurate in this population (17). In all cases, the etiology of kidney dysfunction could not be determined due to the presence of features compatible with CKD, multiple risk factors for CKD or renal dysfunction of unknown duration. Of the 37 patients with GFR<40 mL/min/1.73 m2, 21 patients had proteinuria and/or hematuria, 10 had ≥2 risk factors for CKD and renal failure of unknown duration, 2 had renal cortical scarring on radiological imaging, 2 had renal failure of unknown duration, 1 patient had solitary kidney and 1 patient was previously on RRT and did not completely recover kidney function. Of the 7 patients on RRT at time of biopsy, 2 had proteinuria and ≥2 risk factors for CKD, 2 were on RRT ≥5 weeks, and 3 had renal failure of unknown duration prior to RRT initiation. Liver transplant candidates who had renal dysfunction following a clear precipitating event or in whom normal kidney function was previously documented were considered to have reversible renal disease and were considered for LTA without undergoing kidney biopsy. Similarly, candidates who had been on RRT for ≥ 8 weeks were considered for SLK transplantation without requiring kidney biopsy. Data regarding candidates' demographics, clinical and laboratory findings and posttransplant course were abstracted from patients' charts and electronic medical records.

Kidney biopsy

In all cases, percutaneous kidney biopsy was performed by an interventional radiologist. The lower pole of the kidney was identified with ultrasound in 42 cases while CT guidance was needed in the remaining 2 cases. An 18 gauge automated biopsy needle was used in all 39 cases with available information regarding type and gauge of the biopsy needle. Biopsy needles used included; (i) BioPince (American Medical Instruments Inc., Dartmouth, MA), n = 28; (ii) Bard Max Core (Bard Inc., Covington, GA), n = 8 and (iii) Temno Coaxial needle (Iskus Health Ltd. Dublin, Ireland), n = 3. One to two passes were made and samples were delivered to the pathologist who determined adequacy. Additional samples were obtained as necessary. All patients had a coagulation profile and complete blood count (CBC) determination within 24 h from the biopsy. Patients with an international normalization ratio (INR) >1.5 or platelet count <50 000/mL received fresh frozen plasma or single-donor platelet transfusion as needed to maintain INR ≤1.5 and platelet count ≥50 000/mL, respectively. Following biopsy completion, patients were monitored overnight for complications. Patients laid flat in the bed for the first 6 h. Vital signs were obtained every 15–30 min in the first 4 h and at less frequent intervals afterward. Voided urine was collected in clear containers to monitor for the development and clearance of macroscopic hematuria. Repeat CBC was obtained 4 h after the biopsy and on the subsequent morning or earlier if needed. Additional testing was obtained in the case of hypotension, flank pain or relentless hematuria. Blood transfusion was provided in case of hemodynamically significant bleeding or if the hemoglobin dropped by more than 1 gm/dL compared to prebiopsy values. Postbiopsy complications were defined as follows: (i) Minor complications: drop in hemoglobin by more than 1 gm/dL without hemodynamic instability and the need for <2 units of packed red blood cell (PRBC) transfusion; (ii) Major complications: ≥2 units PRBC transfusion or the need for intervention.

Sampled renal tissue was preserved and examined by an experienced renal pathologist. Specimens from each biopsy were allocated and processed for routine light, immunofluorescence and electron microscopy. For light microscopy, 2 μ sections were stained with hematoxylin-eosin, silver methenamine, PAS and Masson trichrome stains. Adequate material for interpretation was available in all cases. Adequate sample was defined as the presence of at least 15–20 glomeruli.

Kidney Allocation

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Kidney Allocation
  6. Results
  7. Discussion
  8. Disclosure
  9. References

A decision to proceed with SLK transplant was made if the biopsy showed any one of the following criteria: (i) ≥40% global gGS, (ii) ≥30% interstitial fibrosis (IF) and/or (iii) diffuse (involving ≥50% of glomeruli) membranoproliferative glomerulonephritis (MPGN).

Interobserver variation

To determine if interobserver variation could have affected kidney allocation, we conducted a separate analysis in which two experienced renal pathologists (XG and CC) unaware of the original histological findings or the decision of kidney allocation blindly reviewed the biopsy specimen and scored them for the degree of IF and gGS.

Posttransplant management

Posttransplant immunosuppression for SLK recipients included induction with rabbit antithymocyte globulin 1.5 mg/kg/day for a minimum of four doses in hepatitis C virus (HCV) negative patients or basiliximab 20 mg at day 0 and 4 in HCV positive patients. Intravenous methylprednisolone was given for 4 days followed by oral immunosuppression with prednisone tapered to 5 mg/day by 3 months; mycophenolate mofetil 1000 mg twice daily and tacrolimus (target trough level 7–11 ng/mL for 3 months and 4–7 ng/mL thereafter (HPLC-MS/MS whole blood assay). LTA recipients received basiliximab similar to SLK recipients; intravenous methylprednisolone on day 0 followed by tacrolimus, mycophenolate mofetil 1000 mg twice daily for 4 months and prednisone taper to 0 mg/day at 4 months. Conversion from tacrolimus to sirolimus was undertaken in selected patients as felt necessary by the transplant team. Antimicrobial prophylaxis included: (i) valgancyclovir 450–900 mg/day for 3–6 months in all SLK recipients and LTA recipients who received antithymocyte globulin, (ii) trimethoprim/sulfamethoxazole for 6–12 months and (iii) clotrimazole for 1 month for SLK transplants. Intravenous caspofungin was given to critically ill patients. Posttransplant renal function was assessed by serial serum creatinine and iothalamate GFR obtained at 4 months and then yearly. Missing GFR values were replaced by 24-h urinary creatinine clearance when available.

Data analysis

Results are presented as counts and percentages for qualitative data and means and standard deviations for quantitative data. Means of normally distributed data were compared by Student's t-test and data that were not normally distributed were compared by nonparametric tests. Analysis of variance (ANOVA) was used to compare means between more than two groups. Chi-square was used for two by two table associations. Measurement of agreement was analyzed using weighted kappa statistics with the following thresholds: kappa 0.61–0.8 = good and kappa >0.81 = excellent reproducibility. A logistic regression analysis was conducted to identify predictors of biopsy complications. Due to the small number of events, a multivariate analysis could not be performed.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Kidney Allocation
  6. Results
  7. Discussion
  8. Disclosure
  9. References

Demographic information including age, gender, cause of end-stage liver disease (ESLD) and results of kidney function and coagulation parameters at time of kidney biopsy are summarized in Table 1. The study cohort consisted mainly of men with HCV-induced liver disease who were considered for primary liver transplantation. Seven patients (16%) were on RRT for a mean duration of 14.7 days (range 1–39 days) at time of PRB.

Table 1.  Demographic information, laboratory and ultrasound data on 44 liver transplant candidates at time of diagnostic kidney biopsy
  1. Note:1Values are provided only for patients not on RRT at time of kidney biopsy.

Age (years) mean ± SD56.9 ± 7.9
Male gender number (%)31 (71)
Cause of ESLD number (%)
  HCV19 (43)
 Cryptogenic cirrhosis10 (23)
 Alcoholic cirrhosis7 (16)
 Nonalcoholic steatohepatitis (NASH)2 (4)
 Autoimmune hepatitis2 (4)
 Other4 (9)
Primary liver transplant number (%)34 (77)
RRT number (%)7 (16)
Serum creatinine (mg/dL) mean ± SD11.8 ± 0.5
GFR mL/min/1.73 m2 mean ± SD128.4 ± 8.1
24-h protein excretion (mg/day) median (range)174 (0–13 625)
24-h urinary sodium mean ± SD165.3 ± 62.1
FENA <1 number (%)134 (92)
Diuretic use number (%)125 (67)
Mean kidney size (cm) mean ± SD11.3 ± 1.1
INR mean ± SD 1.4 ± 0.2
Platelet count (cells/mL) mean ± SD106 ± 74

For those not on RRT, mean serum creatinine and GFR were 1.8 ± 0.5 mg/dL and 28.4 ± 8.2 mL/min/1.73 m2, respectively. Fractional excretion of sodium (FeNa) of less than 1 was detected in the majority (92%) of patients despite diuretic use in 67% of the cases consistent with an avid renal sodium conservative state. Twenty-one patients (57%) had either proteinuria ≥150 mg/day and/or hematuria; 10 (27%) had both proteinuria and hematuria, 6 (16%) had proteinuria alone and 5 (14%) had hematuria only.

Histological findings on kidney biopsy

Kidney biopsy findings are summarized in Table 2. Positive immunofluorescence staining for immunoglobulin A (IgA) was the most common abnormality occurring in 20 (45%) cases. The majority of these showed trace and segmental mesangial staining, although in four cases staining for IgA was diffuse and global. Since more than one histological abnormality was present in 28 (64%) patients, cases were classified according to the predominant histological diagnosis into: (i) Acute tubular necrosis (ATN), n = 13 (30%); (ii) MPGN, n = 5 (11%); (iii) Advanced IF and/or gGS (IF/GS) defined as IF ≥30% and gGS ≥40%, n = 15 (34%) and (iv) minimal histological abnormalities, n = 11 (25%). Table 3 summarizes the results of prebiopsy evaluation for these four groups. Interestingly, prebiopsy creatinine, GFR and kidney size did not relate to the histological diagnosis (p > 0.05 for all variables) and did not predict reversible renal disease. Mean urinary protein excretion and percentage of patients with hematuria was higher in MPGN cases (p = 0.05 and p = 0.007, respectively). The relationships between the histological diagnosis and prebiopsy GFR and 24-h urinary protein excretion are highlighted in Figures 1 and 2.

Table 2.  Kidney biopsy findings in 44 liver transplant candidates with renal failure
Histological finding1Number (%)
  1. IgAN = IgA nephropathy; MPGN = membranoproliferative glomerulonephritis; FSGS = focal segmental glomerulosclerosis.

  2. 1More than one finding was present in 28 patients.

Glomerulonephritis31 (71)
IgAN20 (45)
MPGN 6 (14)
Diabetic nephropathy 5 (11)
FSGS4 (9)
ATN/injury18 (41)
Advanced IF12 (27)
Advanced gGS 7 (16)
Table 3.  Clinical, laboratory and ultrasound information of 44 liver transplant candidates with renal failure according to the predominant histological diagnosis
 ATN n = 13MPGN n = 5Minimal findings n = 11IF/GS n = 15p
  1. 1Values provided only for patients not on RRT.

  2. ATN = acute tubular necrosis/injury; MPGN = membranoproliferative glomerulonephritis; IF/GS = advanced interstitial fibrosis or glomerulosclerosis.

HCV-induced ESLD number (%)5 (38)5 (100)4 (36)5 (33)0.49
Primary liver transplant number (%)12 (92)4 (80)9 (82)9 (60)0.22
RRT number (%)5 (38)0 (0)0 (0)2 (13)0.04
Serum creatinine (mg/dL) mean ± SD1 1.7 ± 0.5 1.6 ± 0.3 1.6 ± 0.3 2.1 ± 0.60.12
GFR mL/min/1.73 m2 mean ± SD126.2 ± 7.331.2 ± 7.931.4 ± 7.526.1 ± 8.90.30
24-h protein excretion (mg/day) median (range)139 (0–162)522 (119–13  625)78 (0–2464)138 (0–2162)0.05
Proteinuria ≥150 mg/day number (%)11 (12)4 (80)5 (45)6 (46)0.11
Hematuria number (%)10 (0)5 (100)5 (45)5 (38) 0.007
FENA <1 number (%)17 (87)5 (100)10 (91)12 (92)0.88
Mean kidney size (cm) mean ± SD11.6 ± 1.011.3 ± 0.611.2 ± 1.711.1 ± 1.00.76
image

Figure 1. Relationship between kidney biopsy diagnosis and GFR (mL/min/1.73 m2) for 37 liver transplant candidates with renal failure not on RRT. Plot of iothalamate GFR for liver transplant candidates with ATN, MPGN, advanced IF and/or gGS and minimal histological findings. No difference in GFR was observed between the four histological groups (p = 0.30). The line represents the mean GFR of the 37 candidates.

Download figure to PowerPoint

image

Figure 2. Relationship between kidney biopsy diagnosis and 24-h urine protein excretion in 37 liver transplant candidates with renal failure not on RRT. Plot of 24-h urinary protein excretion and biopsy findings. Protein excretion was higher in patients with MPGN (p = 0.05). The line represents the median 24-h protein excretion of the 37 candidates.

Download figure to PowerPoint

Biopsy complications

Thirty-one (70%) candidates did not develop any postbiopsy complications and did not require blood transfusion. Minor complications developed in five (12%) cases. Major complications occurred in eight (18%) cases; one patient developed an arterio-venous fistula that did not require any intervention and seven cases had retroperitoneal hematoma and/or macroscopic hematuria, five of these required fluoroscopic guided selective coil embolization while in the remaining two bleeding stopped spontaneously. No patient required surgical intervention and there were no deaths related to the biopsy. Overall 30% of patients were considered to have had a complication. Table 4 shows the results of logistic regression analysis for predictors of biopsy complications. Only INR ≥ 1.5 predicted biopsy complication (HR = 5.28, 95% CI 1.2–23.2, p = 0.03) Factors that did not relate to biopsy complications included; serum creatinine, RRT at time of biopsy, platelet count and number of passes made. Also, the type of biopsy needle used had no relationship to postbiopsy bleeding (data not shown).

Table 4.  Logistic regression analysis for predictors of PRB complications
 Hazard ratio95% Confidence intervalp
  1. 1Only patients not on RRT were included in this analysis.

RRT2.030.4–10.70.41
Serum creatinine (mg/dL)10.670.1–3.40.62
INR ≥1.55.281.2–23.20.03
Platelet count0.980.9– 10.09
Number of passes1.450.8–2.50.20

Kidney allocation

Based on the kidney biopsy result, 27 (61%) and 17 (39%) candidates were listed for LTA and SLK transplantation, respectively. Table 5 summarizes patients' demographics and results of prebiopsy laboratory, ultrasound and histological information for these patients. More liver retransplant candidates were listed for SLK transplants probably reflecting fixed renal damage from prior calcineurin inhibitor exposure. Similar proportions of candidates who were later listed for either LTA or SLK were receiving RRT at time of biopsy. Duration from RRT initiation to PRB was shorter in SLK candidates (p = 0.05). Of interest, five of the seven cases (71%) on RRT at time of PRB had ATN and hence were listed for LTA. For candidates not on RRT, serum creatinine was higher in those eligible for SLK transplants (2.0 ± 0.6 mg/dL vs. 1.6 ± 0.3 mg/dL, p = 0.02) but GFR was similar between LTA and SLK candidates (30.1 ± 7.8 mL/min/1.73 m2 and 25.9 ± 8.3 mL/min/1.73 m2, respectively, p= 0.12). Mean kidney size, urinary sodium indices, urinary protein excretion, percentage of patients with proteinuria >150 mg/day or with hematuria were also not different between LTA and SLK candidates and did not help with the decision making for kidney allocation. As shown in Table 5, the histological findings were the main determinants of kidney allocation with more IF and gGS in SLK candidates (31.7± 14.8% and 32.3 ± 16.7%, respectively) while LTA candidates had ATN or miminal histological findings. Of the 17 patients listed for SLK, 6 (35%) had IF only, 3 (18%) had gGS only, 5 (29%) had both IF and gGS and 3 (18%) had diffuse MPGN, one of these had concomitant IF. Two patients had MPGN that involved less <50% of the glomeruli and were considered for LTA. One patient who had multiple comorbidities was considered for LTA despite having advanced gGS (involving 69% of the glomeruli). This patient was later deemed not suitable for transplantation.

Table 5.  Demographics, clinical, ultrasound and histological findings at time of kidney biopsy for 44 liver transplant candidates accepted for LTA or SLK
 LTA n = 27SLK n = 17p
  1. 1Values provided only for patients not on RRT.

Age (years) mean ± SD55.9 ± 7.758.5 ± 8.10.30
Male gender number (%)18 (67)13 (76)0.49
HCV-induced ESLD number (%)11 (41)8 (47)0.40
Primary liver transplant number (%)24 (89)10 (59)0.02
RRT at biopsy number (%)5 (18)2 (12)0.55
Duration from RRT to biopsy (days) median (range)10 (7–39)1 (0–1)0.05
Serum creatinine (mg/dL) mean ± SD11.6 ± 0.32.0 ± 0.60.02
GFR mL/min/1.73 m2mean ± SD130.1 ± 7.825.9 ± 8.30.12
24-h protein excretion (mg/day) median (range)170 (0–2464)164 (0–13 625)0.35
Proteinuria ≥150 mg/day number (%)16 (28)5 (33)0.76
Hematuria number (%)18 (38)7 (47)0.61
FENA <1 number (%)120 (91)14 (93)0.79
Diuretic use number (%)114 (64)11 (73)0.54
Ultrasonographic findings
Mean kidney size (cm) mean ± SD11.5 ± 1.211.0 ± 0.90.17
Biopsy findings
Histological diagnosis number (%)
 ATN13 (48)0 (0) 
 MPGN2 (7)3 (18)  0.0001
 IF/GS1 (4)14 (82) 
 Minimal pathological changes11 (41)0 (0) 
Percent gGS (%) mean ± SD 17.1 ± 13.332.3 ± 16.7 0.002
Percent IF (%) mean ± SD11.5 ± 6.431.7 ± 14.8  0.0001

Interobserver variation

Concordance between pathologists on ≥30% IF and ≥40% GS occurred in 88% and 93% of the cases available for review, respectively (pathology slides were missing on one patient). The kappa coefficient was 0.67 (good) for IF ≥30% and 0.8 (excellent) for GS ≥40%. In the five cases with disparity between pathologists, the degree of IF was in the 20–30% range.

Posttransplant course

After a mean of 77 (range 1 – 300) days, 16 patients received LTA and 8 received SLK transplants. In the same time period, another 66 and 17 patients with serum creatinine ≥1.5 mg/dL or on RRT received LTA and SLK transplants, respectively, at our institution without undergoing a kidney biopsy. Characteristics and posttransplant information on the 24 patients who received a transplant based on PRB results are presented in Table 6. SLK candidates waited longer from biopsy to transplantation (128 ± 89 days vs. 52 ± 34 days, p = 0.05) and were more likely to have had a previous liver transplantation. All 8 recipients (5 LTA and 3 SLK) on RRT at time of transplantation recovered kidney function and discontinued dialysis after transplantation. For those not on posttransplant RRT, 1 month serum creatinine was higher in LTA recipients but it was similar to SLK recipients at last follow-up. Eleven and 10 LTA recipients and 7 and 4 SLK recipients had 4-month and 1-year GFR available. Importantly, 4-month and 1-year GFR were comparable between LTA and SLK recipients (Table 6). There was a trend for more rapamune use in LTA patients (31% vs. 0%, p = 0.08) but this did not reach statistical significance.

Table 6.  Characteristics at time of transplantation, posttransplant renal function and immunosuppression for 16 LTA and 8 SLK transplant recipients
 LTA n = 16SLK n = 8p
  1. 1Values provided only for patients not on RRT.

Age at transplant (years) mean ± SD57.6 ± 6.161.2 ± 8.20.24
MELD score mean ± SD23 ± 525 ± 70.44
Time from Biopsy to transplant (days) mean ± SD52 ± 34128 ± 890.05
Primary liver transplant number (%)15 (94)4 (50)0.03
RRT at transplant number (%)5 (31)3 (37)0.76
1-month RTT number (%)2 (12)1 (12)1.00
1-month serum creatinine (mg/dL) (mean ± SD)11.7 ± 0.61.1 ± 0.2 0.006
Immunosuppression number (%)
 Anti-thymocyte globulin1 (6)7 (88) 0.001
 Basiliximab13 (81)1 (12) 0.001
 Tacrolimus11 (69)7 (87)0.32
 Rapamune5 (31)0 (0)0.08
Follow-up days mean ± SD392 ± 210424 ± 2930.76
Serum creatinine at time of death or last follow-up (n = 19) (mg/dL) mean ± SD)11.4 ± 0.41.4 ± 0.20.82
RRT at time of death or last follow-up number (%)3 (19)2 (25)0.72
4-month GFR (mL/min/1.73 m2) mean ± SD139.5 ± 15.1 n = 1144.9 ± 15.6 n = 70.48
1-year GFR (mL/min/1.73 m2) mean ± SD150.6 ± 18.5 n = 1040.8 ± 8.4 n = 40.33

Acute kidney injury (AKI) requiring RRT within 4 month from transplantation developed in three LTA and two SLK recipients. In all cases AKI developed following infectious complications. Cytomegalovirus (CMV) disease was the identified organism in three patients while bacteremia developed in two cases. None of the five recipients who required posttransplant RRT recovered kidney function. For the 24 patients who received a transplant, posttransplant mortality was 21% (5,24) after a 403 ± 135 follow-up period. Mortality was higher in those who developed posttransplant RRT-dependent AKI (80% vs. 5%, p = 0.0001).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Kidney Allocation
  6. Results
  7. Discussion
  8. Disclosure
  9. References

This study demonstrates the feasibility, benefit and risk of PRB in 44 liver transplant candidates with moderate-to-severe renal dysfunction of undetermined etiology in whom it was difficult to decide on either SLK or LTA with the available clinical information. Our results demonstrate the discordance between clinical, laboratory and ultrasound information and the pathological diagnosis. Biopsy complications were not uncommon occurring in 30% of cases with major complications occurring in 18% but INR ≥ 1.5 was the single predictor of postbiopsy bleeding. Perhaps most importantly, relying on histological information allowed us to avoid SLK transplantation in 27 (61%) of cases and LTA recipients had favorable posttransplant renal function. All five patients on RRT at time of biopsy who later received LTA recovered kidney function within 1 month from transplantation and post-LTA kidney function was comparable to those who received SLK transplants. Interobserver agreement in interpreting the degree of IF and/or gGS was good-to-excellent indicating the reproducibility of utilizing histological data in deciding on SLK transplantation. This study also highlights the unpredictability of postliver transplant AKI regardless of the biopsy findings mainly due to the increased risk of infections in liver transplant candidates with pretransplant renal dysfunction.

Our results extend previous experiences in utilizing kidney biopsy in liver transplant candidates with renal failure (13–15). Pichler and colleagues performed both transjuglar renal biopsy (TJRB) and PRB on 28 liver transplant candidates and used histological information to recommend SLK transplantation in only 10 (36%) patients with ≥30% IF and/or ≥40% GS (15). Jouet et al. performed TJRB on 15 liver transplant candidates and recommended SLK transplants in five (33%) cases (13). Importantly, clinical criteria commonly used to determine kidney function and to assess chronicity of the renal disease did not relate to histological evidence of fixed renal damage and did not identify potentially reversible renal failure. This lack of correlation is in agreement with previous reports that document a discrepancy between renal biopsy findings and clinical presentations especially in patients with HCV-induced liver disease (13,18). This lack of agreement might also explain why preliver transplant GFR does not consistently predict postliver transplant renal function (19–21). Cohen and colleagues identified pretransplant GFR <40 mL/min/1.73 m2 in 20 LTA recipients, 60% of those had a posttransplant GFR >40 mL/min/1.73 m2 while 3 (15%) developed end-stage-renal disease (ESRD) (20).

In this study, we used ≥30% IF and/or ≥40% gGS as histological criteria to define irreversible kidney damage and to determine the need for combined organ transplantation. These cut-offs are based on native kidney biopsy studies that document the negative impact of advanced IF and gGS on renal prognosis and on previously published reports that suggest ≥30% IF and/or ≥40% gGS as an indication for SLK transplantation (8,12,22–24). In this study, we did not include a control group and hence we cannot establish if patients with similar or more advanced degree of IF/GS would have had acceptable kidney function following LTA. Our findings also do not demonstrate that histological information is more predictive than clinical judgment for kidney allocation to all liver transplant candidates with renal failure since we selected a group of liver transplant candidates with overlapping features of both CKD and hepatorenal physiology in whom the etiology and potential reversibility of the renal failure could not be made on clinical grounds. Our results, however, highlight the discordance between the histological diagnosis and the pretransplant clinical information in this challenging group of liver transplant candidates and suggest that PRB is a safe and reproducible procedure that might be helpful in avoiding unnecessary SLK transplantation. By utilizing histological information, we identified 27 (61%) candidates who did not need SLK transplants including five patients on RRT with biopsy evidence of ATN who received LTA and recovered kidney function within 1 month from transplantation. LTA recipients had also acceptable kidney function with comparable serum creatinine at last follow-up, 4-month and 1-year GFR between LTA and SLK recipients (Table 6). Although these results are encouraging, this single-center experience should be considered as a platform for future randomized trials to determine if indeed PRB can guide kidney allocation in this selected group of liver transplant candidates.

PRB is an established method for sampling the renal parenchyma and is associated with bleeding complications in 3% to 13% of cases (25,26). Whittier and colleagues reported their experience with 750 PRB in noncirrhotic patients and noticed complications in 98 (13%) patients with major complications requiring intervention and/or blood transfusion occurring in 6.4% (25). The current results indicate that postbiopsy complications in cirrhotic patients are not uncommon occurring in 30% of cases with 11% (5,8) requiring radiological intervention to stop the bleeding. Although this might be considered a high complication rate, it should be noted that our patients had a GFR <40 mL/min/1.73 m2, a risk factor on its own for excessive postbiopsy bleeding. Winkelmayer et al. reported 6-folds increased risk of postoperative bleeding in patients with similar degree of renal failure (27). Similarly, native kidney biopsy studies document a 2.3-folds increased risk of postbiopsy bleeding in patients with advanced renal failure (25,28). TJRB is another alternative for renal sampling especially in cirrhotic patients (13–15). Sam and colleagues reported their experience with 29 TJRB in patients with advanced liver disease (14). Complications overall were still common with 28% requiring blood transfusion and 17% developing contrast nephropathy. Pichler et al. performed PRB in eight liver transplant candidates with renal failure of unclear etiology and reported 63% risk of perinephric hematoma and 12.5% risk of blood transfusion/selective embolization (15). Another 23 patients have undergone a total of 25 TJRB. Complications rate was lower with TJRB (4%) but 30% of patients did not have adequate diagnostic sample (15). Also, TJRB carries up to 90% risk of inadvertent capsular perforation and in 10% of cases there is a need for rebiopsy due to the limited sample size (29). Our findings confirm that PRB is feasible in liver transplant candidates when a systematic approach aiming at correcting the underlying coagulopathy is used. Based on our experience, we think that PRB is preferable to TJRB due to the adequate sample size, lack of risk of contrast nephropathy and limited chance of inadvertent bowel perforation. We, however, believe that specific center expertise and comfort level will dictate the optimal biopsy method.

One inherent problem of kidney biopsy is the small sample size that can limit histological interpretation. We were able to obtain an adequate specimen in all our patients as an experienced renal pathologist assessed the specimen during the biopsy procedure and asked for additional samples if needed. This practice obviated the need for rebiopsy and was not associated with increased risk of bleeding as demonstrated in our analysis. Interobserver variation in interpreting the degree of IF or gGS might also limit the usefulness of histological information for kidney allocation. Our results indicate that the interobserver variation was minimal even when only a common method like light microscopy was used. These findings demonstrate that ≥30% IF and ≥40% gGS cut-offs are reproducible and can be applied to daily clinical practice.

Stating that histological evidence of reversible renal dysfunction will eliminate the risk of postliver transplant AKI or RRT is an oversimplification. Despite using histological evidence of reversible renal damage our data showed a persistent risk of posttransplant AKI that might require RRT and might progress to ESRD even in those who received SLK transplantation. This finding is unsurprising and is attributed to the multitude of postliver transplant events that precipitates AKI, of these, infectious complications seem to be the most detrimental (30). Previous reports have documented an increased risk of infections in both LTA and SLK transplant recipients with pretransplantation renal failure compared to other types of kidney or liver transplantation (4,31,32). The negative impact of post-LTA or SLK transplant AKI that requires RRT on posttransplant survival observed in our study is consistent with previous reports (33).

Our study represents the largest single-center experience in using PRB findings to allocate kidneys to liver transplant candidates with renal failure. The study included consecutive patients that might have limited the selection bias inherent to its retrospective design. It should be recognized that our data is limited to the studied population that consisted mainly of patients with moderate-to-severe renal dysfunction of undetermined etiology, who had no clear precipitating event and were on RRT for less than 8 weeks. The criteria we used to list candidates for SLK transplantation were based on extensive IF, gGS and/or diffuse MPGN. Whether candidates with similar or more advanced degree of IF and/or GS can still benefit from LTA transplant is an important question that can not be answered in this study and warrants future investigation.

In conclusion, this study demonstrates that PRB is feasible in liver transplant candidates with moderate-to-severe renal dysfunction of unknown etiology and provides valuable and reproducible histological information that can not be predicted from the pretransplant clinical data. Histological finding of potentially reversible renal damage does not preclude postliver transplant AKI that necessitates RRT primarily due to the increased risk of infection in liver transplant recipients with pretransplant renal failure. Complications from PRB occur in 18–30% of cases and are related to INR ≥1.5. Future randomized studies are needed to confirm that PRB is useful in selecting SLK candidates.

Disclosure

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Kidney Allocation
  6. Results
  7. Discussion
  8. Disclosure
  9. References

This work was accepted in part as a free communication at the American Transplant Congress meeting, Toronto, Canada, 2008.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Kidney Allocation
  6. Results
  7. Discussion
  8. Disclosure
  9. References