Predicting Renal Failure After Liver Transplantation From Measured Glomerular Filtration Rate: Review of up to 15 Years of Follow-Up

Abstract

Background. 

The immunosuppressive medications that have contributed greatly to the success of liver transplantation are also associated with posttransplant renal dysfunction. We reviewed measured glomerular filtration rate (GFR) data from patients who underwent transplantation more than 10 years ago to assess whether results from specific time points can predict renal failure.

Methods. 

The GFR data were obtained at initial evaluation (IE), at month 3, and at years 1, 2, 5, 10, and 15. Two groupings were compared, one based on GFR at IE and the other at month 3. Patients were further stratified into three GFR (mL/min/1.73 m²) groups: G1, GFR more than 80; G2, GFR 60 to 80; and G3, GFR less than 60.

Results. 

A total of 592 liver transplant recipients met the inclusion criteria; 114 had paired GFR data from IE to year 15. Analysis of paired and censored data based on IE GFR showed that 62.2% of G3 patients developed renal failure by year 5; another 6.7% did so by year 10 (P=0.027). The month 3 GFR data showed that 56.3% of G3 patients developed renal failure by year 5; another 15.6% did so by year 10. Surprisingly, 37.0% of G2 patients experienced renal failure by year 5; another 11.1% did so by year 10 (P=0.0024).

Conclusions. 

The month 3 data indicate a slow but steady decline in GFR over years. The lower the initial GFR is after transplant, the sooner renal failure develops. Patients with GFR less than 60 mL/min per 1.73 m² at month 3 have a higher risk of renal failure; however, those who avoid renal failure seem to maintain renal function long term.

The current success of liver transplantation would never have been imagined when the first transplants were completed. This success has largely been due to improvements in surgical technique, in anesthesia, and, ultimately, in the effectiveness of immunosuppressive medications, namely the calcineurin inhibitors (CNIs). Gonwa et al. (¹) were among the first to describe a serious impact of cyclosporine on glomerular filtration rate (GFR) after liver transplantation by using the iothalamate sodium I-125 (Glofil) method. This landmark work eventually described stability in measured GFR between years 1 and 4. Subsequent findings from other centers addressing the renal function issue demonstrated a progressive decline after liver transplantation, without a stability period. A majority of these studies, however, relied only on calculated clearances. As is well known, there are many formulas for calculating creatinine clearance, and none of them are accurate for patients in end-stage liver disease.

The present study, which is a follow-up to the 1995 study by Gonwa et al., is a single-institution experience with measured GFR in our population of liver transplant recipients. The aim is to determine whether measured GFR can be used to predict renal failure in these patients.

MATERIALS AND METHODS

A review of the prospectively maintained research database of all liver transplant recipients at our institution was conducted with the approval of the Baylor Health Care System Institutional Review Board. Data from patients who underwent transplantation between 1985 and 1998 were reviewed to allow for a follow-up period of more than 10 years. Exclusion criteria included renal replacement therapy at time of transplant, combined liver-kidney transplant, fulminant hepatic failure, and patient death within 1 year. Otherwise, all GFR data obtained from this patient population were used for statistical review.

The cyclosporine-based immunosuppressive regimen used at our institution was changed to a tacrolimus-based regimen in 1994. Additionally, azathioprine (used from 1980 to 1990) was replaced by mycophenolate mofetil as the antimetabolite agent in our protocol. Steroid taper was also included in our immunosuppressive regimen for all patients. In cases of acute renal failure after transplantation, treatment with OKT3 was used to delay the introduction of CNI. Acute cellular rejection was treated with steroid recycle, and steroid-resistant rejection was treated with antibody (OKT3 or rat antithymocyte globulin).

GFR data were obtained from protocol for the following time points: at time of initial evaluation (IE), at month 3, and at years 1, 2, 5, 10, and 15. Data were censored for individuals who eventually developed renal failure. In this population, renal failure was defined as (1) requiring renal replacement therapy (hemodialysis or peritoneal dialysis; Kidney Disease Outcomes Quality Initiative stage 5 chronic kidney disease [CKD 5]); (2) being listed for a renal transplant (CKD 5); (3) having GFR less than 30 mL/min per 1.73 m² (CKD 4); or (4) receiving a renal transplant.

Two GFR comparison groupings were used, one based on the value at IE and the other at month 3. These time points were selected to determine their suitability in predicting long-term renal function after liver transplantation. Patients were further stratified into three groups according to GFR (mL/min/1.73 m²): G1, GFR more than 80; G2, GFR 60 to 80; and G3, GFR less than 60. Paired data analysis (Wilcoxon two-sample test) was then used to compare the two time point groupings, IE and month 3; P less than 0.05 was considered significant. Data were censored for the development of renal failure and death.

RESULTS

From the review of the research database, 592 liver transplant recipients met the inclusion criteria, and 114 of them had paired GFR data from IE to year 15. The mean age of the recipients was 49.0 years, and the mean graft survival time was 121.7 months (range, 12.5–267.9 months). A cyclosporine-based regimen was used for 374 patients (63.4%), and 195 (32.9%) patients were given a tacrolimus-based regimen. For the patients in the overall study population (nonpaired data), the development of renal failure during the 15-year follow-up is summarized in Table 1, along with details such as dialysis, CKD 4 classification, kidney transplantation, and death.

TABLE 1: Renal function and kidney transplant status of patients in the overall study population according to year of follow-up after liver transplantation

Overall Nonpaired, Noncensored GFR Data

Analysis of the overall data indicated stability, and the GFR distribution over time indicated preservation of renal function, with GFR more than 50 mL/min per 1.73 m² at 15 years after liver transplantation. No predictive significance was found when the nonpaired data were used to evaluate the decline in slope between any two time points (P>0.05) or the actual degree of reduction in GFR between any two time points (P>0.05; Wilcoxon two-sample test).

Comparison of Time Point Groupings

In the grouping based on IE GFR, the decline in mean GFR was more pronounced for G1 than for G2. Patients in G3 had an increase in mean GFR by month 3. All three groups showed a stable GFR profile until year 15 (Fig. 1). The highest incidence of renal failure by year 5 was in G3 (62.2% of patients), and an additional 6.7% of this group developed renal failure by year 10 (P=0.027; Fig. 2).

FIGURE 1.: Mean measured glomerular filtration rates (GFRs) after the initial evaluation (IE) GFR was used to group patients into G1 (GFR>80 mL/min/1.73 m²), G2 (GFR 60–80 mL/min/1.73 m²), and G3 (GFR<60 mL/min/1.73 m²). Data paired and censored.

FIGURE 2.: Development of renal failure (%) based on grouping by initial evaluation (IE) glomerular filtration rate (GFR; Wilcoxon two-sample test, P=0.027).

In the grouping based on month 3 GFR, a progressive decline in mean GFR occurred in G1 and G2 (Fig. 3). Analysis showed that 56.3% of patients in G3 developed renal failure by year 5, with an additional 15.6% developing renal failure by year 10. Surprisingly, 37.0% of patients in G2 developed renal failure by year 5, and an additional 11.1% developed renal failure by year 10 (P=0.0024; Fig. 4).

FIGURE 3.: Mean measured glomerular filtration rates (GFRs) after the month 3 GFR was used to group patients into G1 (GFR>80 mL/min/1.73 m²), G2 (GFR 60–80 mL/min/1.73 m²), and G3 (GFR<60 mL/min/1.73 m²). Data paired and censored.

FIGURE 4.: Development of renal failure (%) based on grouping by month 3 glomerular filtration rate (GFR; Wilcoxon two-sample test, P=0.0024).

Immunosuppression Effect

Data were analyzed on the basis of immunosuppressive agents being used at month 3 and year 1 after liver transplantation. The comparison between tacrolimus, cyclosporine, and sirolimus did not demonstrate a significant difference in the month-3 (M3) GFR. Tacrolimus demonstrated better preservation of long-term renal function with statistical significance (P<0.05) when used at M3 and year 1. However, this analysis was subject to selection bias at time points greater than 5 years after liver transplantation (Table 2).

TABLE 2: Median GFR based on immunosuppression agent at M3 and Y1

DISCUSSION

Ojo et al. (²) in 2003 reported chronic renal failure in 18.1% of recipients at 5 years after liver transplantation. Their analysis showed that relative risk increases to 2.54 when the calculated pretransplant GFR is less than 59 mL/min per 1.73 m². The relative risk increases to 3.78 when the calculated pretransplant GFR is less than 29 mL/min per 1.73 m². Other studies show an 18.1% incidence of severe renal dysfunction (9.5% incidence of end-stage renal disease) 13 years after liver transplantation (³).

Additional studies have demonstrated progressive decline in renal function after liver transplantation. Pawarode et al. used estimated GFR (from the Modification of Diet in Renal Disease Study equation) and found that 35% of the liver transplant patients they studied had developed permanent renal dysfunction. The independent predictors included serum creatinine more than 1.2 mg/dL and a baseline GFR less than 70 mL/min per 1.73 m². A decrease in GFR of more than 30 mL/min per 1.73 m² at 9 months predicted development of permanent renal dysfunction, and absolute GFR less than 30 mL/min per 1.73 m² occurring as early as 3 months posttransplant predicted severe renal failure (⁴). Others using calculated or estimated GFR have reported similar findings (⁵).

Cohen et al. (⁶) used measured GFR, with results similar to ours. They found that the 1-year GFR correlated better with the 3-year GFR than with the pretransplant GFR and concluded that pretransplant GFR did not seem to be a good marker for identifying patients at risk for late renal dysfunction. Those with pretransplant GFR less than 60 mL/min per 1.73 m² developed significant renal dysfunction by 3 years.

We used only measured protocol GFR to determine renal function after liver transplantation rather than calculated/estimated GFR. Analysis of the equations for calculating renal function after liver transplantation has shown that there is no single best method (⁷). The most important features of the present study are the paired data, censoring for chronic renal failure, and the long-term (15-year) follow-up. This study demonstrates what seems to be the true chronological history of renal function after liver transplantation. What it does not reflect, however, are the immunosuppression practices that could have altered the data points at any time during the follow-up (mainly, minimizing the effects of CNI by adding an antimetabolite such as azathioprine or mycophenolate mofetil, discontinuing CNI, and/or initiating sirolimus to delay or reverse the renal dysfunction resulting from CNI use). In the limited analysis looking specifically at immunosuppression types at M3 and year 1, tacrolimus has better renal preservation long-term up to 10 years when compared with cyclosporine. However, these data are subject to selection bias because of practice patterns of immunosuppression minimization to preserve renal function as mentioned earlier. It also seems that sirolimus has poor long-term renal function preservation; however, the population number is small and there was a tendency to convert to sirolimus in circumstances where there was extremely poor renal function. At this time, there is not a significant number of patients to perform long-term (>10 years) paired data analysis.

Three conclusions can be drawn from this review. First, patients with GFR less than 60 mL/min per 1.73 m² (CKD 3) at month 3 after liver transplantation have a higher risk of developing renal failure. Second, the month 3 data indicate a slow but steady decline in GFR over a period of years after liver transplantation. The lower the initial GFR is after transplant, the sooner the renal failure develops (before year 5). Third, and most interestingly, even though patients with GFR less than 60 mL/min per 1.73 m² at month 3 have a higher rate of renal failure, those who do not develop renal failure seem to maintain their renal function long term. Although this last conclusion is based on a smaller number of patients because of the nature of the paired data and losing individuals to follow-up, analyzing this group further may lead to improved management principles.

It is not clear whether these patients with M3 GFR less than 60 mL/min per 1.73 m² will benefit from combined liver kidney transplant because the M3 GFR cannot be predicted before the time of transplant. Further analysis looking specifically at development of end-stage renal disease by the first year after liver transplantation could develop a clearer picture to help answer the question. Interestingly, the preliminary analysis comparing immunosuppressive types suggests that tacrolimus may have better long-term renal preservation after liver transplantation. Long-term censored paired data analysis must be performed to develop a more precise analysis.

Present studies are aimed at determining the effect of specific immunosuppressive agents and regimens on renal function after liver transplantation. This review will also be broadened to include additional nonstandard time points to increase the number of paired data points in the analysis.