The Network Center for Biomedical Research in Hepatic and Digestive Diseases is funded by the Carlos III Institute of Health. Marina Berenguer received a grant (PI-050981) from the Carlos III Institute of Health.
The authors have no conflicts of interest to disclose.
End-stage liver disease due to chronic hepatitis C virus (HCV) infection is the most common indication for liver transplantation (LT) in Europe and in the United States. Unfortunately, HCV recurrence is universal, leads to recurrent disease in many, and ultimately limits long-term graft and patient survival.[1, 2] Although antiviral therapy with pegylated interferon (PEG-IFN) and ribavirin (RBV) is challenging because of (1) the frequent comorbidities that affect patients' adherence and tolerance, (2) safety issues, and (3) hyporesponsiveness, it results in a sustained virological response (SVR) in approximately 30% of treated recipients.[1-6] On the basis of studies of the immunocompetent population as well as preliminary data for LT patients,[7, 8] increased rates of viral clearance are expected with triple therapy.
Renal impairment is a frequent complication in LT recipients and is often associated with reduced survival and increased morbidity. The etiology of chronic kidney disease (CKD) after LT is often attributed to calcineurin inhibitor (CNI) toxicity, but additional etiologies, including hypertension, diabetes, patient age, and HCV infection, need to be considered.[9-11]
The Kidney Disease Outcomes Quality Initiative/Kidney Disease: Improving Global Outcomes classification of kidney disease stratifies renal function into 5 stages. Stage 1 includes patients with a normal or high glomerular filtration rate (GFR) ≥ 90 mL/minute/1.73 m2. Stage 2 includes patients with evidence of kidney damage and a GFR of 60 to 89 mL/minute/1.73 m2. Stages 3 to 5 are defined by reductions in GFR (stage 3, 30-59 mL/minute/1.73 m2; stage 4, 15-29 mL/minute/1.73 m2; and stage 5, <15 mL/minute/1.73 m2). Patients with at least stage 3 CKD have an increased risk of all-cause mortality, cardiovascular mortality, and progressive CKD.[12, 13] In clinical practice, GFR is difficult to measure. Gold-standard methods that measure GFR precisely, such as inulin and validated isotope clearances, are costly and time-consuming. GFR is hence normally calculated via equations using serum creatinine and/or serum cystatin, race, sex, and age. A 4-variable equation that was developed in the Modification of Diet in Renal Disease (MDRD) study to determine kidney function [the 4-variable Modification of Diet in Renal Disease (MDRD-4) equation] is now widely used to estimate GFR in CKD. These equations have been evaluated in a large cohort of LT recipients.
There is increasing evidence for an association between HCV infection and kidney dysfunction/disease. Indeed, recent epidemiological studies strongly suggest that chronic HCV infection is a risk factor for proteinuria and/or impaired renal function in both nontransplant patients and kidney transplant/LT recipients.[11, 17] On the basis of this potential association, we hypothesized that the sustained virological clearance of an HCV infection would result in an improvement in renal function in the LT setting.
We thus aimed to determine in a cohort of LT recipients treated for HCV with PEG-IFN/RBV whether SVR was associated with an improvement in renal function. The hypothesis was that viral eradication would result in an improvement in renal function in patients with an altered GFR before the initiation of antiviral therapy.
PATIENTS AND METHODS
The main study group comprised all LT patients who had undergone transplantation between 1992 and 2009, who had been treated at our institution with PEG-IFN–based therapies for established recurrent HCV disease, for whom the type of response to antiviral therapy had been established (ie, SVR versus nonresponse), and who had a minimum follow-up of at least 1 year since the end of treatment (EOT). For patients who had been retreated, information was collected for the most recent regimen. Indications for antiviral therapy have changed over the years, but they have generally involved progressive disease, fibrosis (>stage 1), or both.
All patients were examined at least every 2 to 6 months. Liver and kidney function tests were performed on each of these visits. Abdominal ultrasound imaging was performed every 6 to 12 months or when it was clinically indicated. The reported GFR levels were reflective of a sustained GFR range averaged over a period of time (greater than at least 3 months), and they hence did not reflect a single time point.
In order to determine whether there were changes in renal function over time, changes in MDRD-4 values 1, 3, and 5 years after therapy were examined as a function of the pretreatment CKD stage.
In addition, in the overall treated population, we determined the prevalence of clinically relevant renal dysfunction (RD), which was defined as an MDRD-4 value < 60 mL/minute at the last follow-up, and we evaluated factors associated with this outcome, including the potential effect of viral clearance. In this second part of the study, we chose the cutoff of 60 mL/minute because this was the cutoff previously associated with progressive CKD, all-cause mortality, and cardiovascular mortality.[12, 13] In those who died or lost their graft, data regarding renal function at the last follow-up were extracted before there was evidence of acute deterioration in kidney function associated with terminal diseases. The follow-up was terminated at the time of the patient's death or at the end of the observation period (December 2012). The study protocol received a priori approval by the the hospital and liver transplantation review committee.
Antiviral Treatment Regimen
Antiviral therapy consisted of PEG-IFN (Pegasys from Roche, Inc., or PegIntron from Schering-Plough, Inc.) in combination with RBV (Rebetol from Schering-Plough or Copegus from Roche) for an intended duration of therapy of 48 weeks whenever possible.
A virological response was defined as negativity for HCV RNA in serum according to a qualitative polymerase chain reaction. The response was considered complete if it occurred at the completion of therapy. An SVR was considered when the virological response was observed 6 months after the completion of therapy.
Factors Predictive of Outcomes
Analyzed variables potentially influencing renal function at the last follow-up included the following: SVR versus nonresponse, sex, age at treatment and at the last follow-up, donor age, time since EOT, pretreatment and EOT MDRD-4 values, CNI trough levels at EOT and 1 year after treatment, diabetes before and after treatment, arterial hypertension, and fibrosis at the baseline and at the last follow-up (when these data were available).
Changes in renal function at different time intervals since EOT were evaluated for the different CKD categories. In addition, in the global cohort of treated patients, the primary endpoint of the analysis was RD at the last follow-up, which was defined as an MDRD-4 value < 60 mL/minute. Baseline characteristics were described as proportions or as medians and ranges. Categorical data were compared with a χ test or Fisher's exact test when indicated. When categorical variables were ordered, comparisons were made with a χ test for trends. Continuous variables were expressed as medians and ranges and were compared with the Mann-Whitney test. Variables with a P value < 0.1 in the univariate analysis were entered into a multivariate regression analysis. A P value < 0.05 was considered to be significant.
One hundred seventy-five treated patients with at least 1 year of follow-up since EOT were included in this study (Table 1). Most were men (74%) with a median age at the last follow-up of 60.5 years (range = 35-79 years). Seventy-five patients achieved an SVR (43%). The median MDRD-4 values at EOT and at the end of follow-up were 73.7 and 70.9 mL/minute, respectively. The median duration of follow-up from EOT until the determination of renal function was 3.8 years (range = 1-9 years).
Table 1. Pretreatment Demographics of the Study Patients Grouped by CKD Stages
All Patients (n = 175)
Stage 1 CKD (n = 45)
Stage 2 CKD (n = 99)
Stage 3 CKD (n = 31)
NOTE: There were no patients with stage 4 or 5 CKD. Percentages for the 3 CKD stage groups are based on the n values for all patients.
The majority of the patients were classified as stage 2 CKD (n = 99), whereas the rest of the categories had significantly fewer patients (45 patients with stage 1 CKD, 31 patients with stage 3, and no patients with stages 4 or 5). The characteristics of the patients belonging to these 3 CKD categories are shown in Table 1.
Evolution of Renal Function in All Categories
Changes in renal function over time as a function of the pretreatment CKD stage are shown in Figs. 1 and 2. There were 45 patients with a high pretreatment GFR (>90 mL/minute), and worsening of renal function over time occurred in only 10 of these patients [6 in the SVR group and 4 in the nonresponder (NR) group].
In addition, no significant changes in renal function after antiviral therapy were observed in the 31 patients with stage 3 CKD at the time of therapy, regardless of their responses to antiviral therapy. For SVR patients, the estimated glomerular filtration rates (eGFRs) before treatment and 5 years after treatment were 54 ± 5 and 52 ±15 mL/minute, respectively, whereas for NRs, these values were 52 ± 6 and 54 ± 15 mL/minute, respectively (P = not significant).
Patients belonging to stage 2, the largest CKD category (n = 99), were the ones in whom changes in MDRD values were observed over time, and there were significant differences between SVR patients and NRs (Fig. 2). The median MDRD-4 values before treatment and 1, 3, and 5 years after treatment were 73.08, 70.5, 79.5 amd 82.1 mL/minute, respectively, for SVR patients and 75.01, 72, 73.9 and 66.1 mL/minute, respectively, for NRs (P = 0.17, P = 0.62, P = 0.16 and P = 0.005, respectively). The median MDRD-4 values before treatment and 1, 3, and 5 years after treatment were 73.08, 70.5, 79.5 amd 82.1 mL/minute, respectively, for SVR patients and 75.01, 72, 73.9 and 66.1 mL/minute, respectively, for NRs (P = 0.17, P = 0.62, P = 0.16 and P = 0.005, respectively).
In order to determine potential causes that could explain the improvement in renal function in SVR patients and, in contrast, the slight deterioration in renal function in NRs among patients with stage 2 CKD before treatment, we compared SVR patients to NRs in terms of other risk factors for CKD, such as age, diabetes, advanced liver disease, and posttreatment modifications of CNIs [ie, levels of cyclosporine A (CSA) and tacrolimus (TAC) at yearly intervals since EOT and percentages of patients who discontinued CNIs since EOT; Table 2]. No differences were found between the 2 groups in any of these variables except for advanced liver disease. Indeed, the proportions of patients with stage 2 CKD before treatment (MDRD-4 value = 60-89 mL/minute) and cirrhosis at the last follow-up were 22% for SVR patients and 56% for NRs (P = 0.002).
Table 2. Comparison of SVR Patients and NRs With Stage 2 CKD
SVR Group (n = 43)
NR Group (n = 56)
The data are presented as medians and ranges.
The data are presented as medians and ranges (with percentages following the slashes).
Finally, because in this study the analysis was based on the initiation of HCV treatment after LT and not all patients started on therapy at the same time, there was the potential for a lead-time bias effect to be created. In order to clarify whether there was a lead-time bias effect, we plotted MDRD versus the time after LT (months) as a function of the viral response to antiviral therapy both for the overall treated population and for the patients with stage 2 CKD before treatment (Fig. 2). Before therapy was started, similar trends in renal function were observed for SVR patients and NRs. In contrast, during and after therapy, we found a progressive but slow decline in eGFR for NRs and a progressive but slow increase in eGFR for SVR patients.
RD at the Last Follow-Up in the Overall Treated Population (n = 175)
Table 3 shows the features of patients with renal impairment (<60 mL/minute) at the last follow-up and patients without RD. RD (MDRD-4 value < 60 mL/minute/1.73 m2) at the last follow-up was present in 54 patients of the overall treated population (31%). RD at the last follow-up was significantly more frequent in NRs versus patients who had achieved an SVR (40% versus 19%, P = 0.002). The post-EOT follow-up did not differ between NRs and SVR patients [3.7 years (range = 1-9 years) versus 4 years (range = 1-8.6 years), P = 0.20].
Table 3. Comparison of Patients With RD at the Last Follow-Up and Patients Without RD (n = 175)
Factors significantly associated with RD at the last follow-up included a lack of an SVR, older age, female sex, and RD before and at EOT. In the multivariate analysis, RD (MDRD-4 < 60 mL/minute/1.73 m2) at the last follow-up was associated with NRs [relative risk (RR) = 3.8, 95% confidence interval (CI) = 1.3-11.23, P = 0.01], EOT MDRD-4 values (RR = 1.022, 95% CI = 1.001-1.04, P = 0.04), and female sex (RR = 5.6, 95% CI = 1.84-17.5, P = 0.002).
Renal impairment is a frequent complication in LT recipients and is often associated with reduced survival and increased morbidity. Half of LT recipients present with RD of a variable grade, and this complication increases with longer follow-up.[17, 18] A large study (n = 36,849) showed a 5-year cumulative incidence of 18% for chronic RD (defined as a GFR < 29 mL/minute/m2). Rates of significant CKD, however, vary largely in published studies.[11, 15, 19] This variability is most likely influenced by the methods used to define this complication, the length of follow-up, the criteria used for patient selection, and the time when transplantation was performed. CNI therapy has been largely considered to be the major factor responsible for the onset of CKD. Additional risk factors associated with posttransplant CKD include preexisting renal insufficiency; the duration of RD in the pretransplant period; perioperative hemodynamic insults to the kidneys; the use of other nephrotoxic drugs; and the presence of hypertension, diabetes, and HCV recurrence.[10, 11, 18-21] Because HCV infection can be eradicated with antiviral therapy, we sought to evaluate whether viral eradication would be associated with improvements in renal function. In order to do so, we retrospectively analyzed the LT database of HCV patients treated with antiviral therapy, and we studied whether viral eradication was followed by changes in MDRD-4 values in those with at least 1 year of follow-up since EOT. Furthermore, we determined the prevalence of significant CKD at the last follow-up in this cohort of patients treated for HCV, and we analyzed factors associated with this outcome. We chose to define significant CKD as an MDRD-4 value < 60 mL/minute because this cutoff identifies patients with at least stage 3 or moderate CKD, and these are the patients with an increased risk for progressive CKD, all-cause mortality, and cardiovascular mortality.[12, 13] Furthermore, there were no patients with stage 4 or 5 CKD, and only 31 patients (18%) had stage 3 CKD at the time of therapy. No significant changes in renal function after antiviral therapy were observed in these 31 patients; for SVR patients, the eGFR values before treatment and 5 years after treatment were 54 ± 5 and 52 ±15 mL/minute, respectively, whereas for NRs, these values were 52 ± 6 and 54 ± 15 mL/minute, respectively. Whether this lack of effect of viral clearance on renal function was due to the small sample size or the renal damage was too evolved to improve over time in this subset of patients with more advanced renal disease cannot be answered with the present data.
The main findings of our study can be summarized as follows:
An improvement in renal function was observed only in patients with a mild degree of CKD (stage 2 CKD) before treatment who achieved an SVR; in contrast, renal function very slowly deteriorated in NRs.
Because no significant differences in CNI use and/or doses were observed between SVR patients and NRs after EOT, the improvement in renal function in stage 2 patients achieving an SVR was attributed to viral eradication among other potential evaluated and unevaluated variables.
In order to determine which factors could have contributed to the changes in renal function, we assessed the overall prevalence of RD in all treated patients; approximately one-third of the patients had at least stage 3 CKD at a median of 4 years after EOT.
Factors independently associated with stage 3 CKD included a lack of viral clearance, female sex, and the presence of previously diagnosed CKD.
HCV is known to be strongly associated with glomerular lesions leading to CKD in both immunocompetent patients and LT recipients. In a US study comparing 18,002 HCV-infected patients to 25,137 controls with an eGFR > 60 mL/minute/1.73 m2, HCV infection was found to be associated with a higher risk for and a shorter time to the development of stage 3-5 CKD. In another comparative study including 512 anti-HCV–positive patients and 313 anti-HCV–negative patients, stage 3-5 CKD was more common in the former group versus the latter group (9.6% versus 5.1%), with HCV-positive patients also showing a shorter time to the development of CKD. Interestingly, a higher baseline viral load was an independent predictor of CKD.
The pathogenesis of renal injury in these patients is thought to be mainly related to the deposition of immune complexes containing HCV proteins, anti-HCV antibodies, and rheumatoid factor immunoglobulin M with anti–immunoglobulin G activity; additional mechanisms are probably involved. Because of the link between HCV infection and renal disease, antiviral therapy has been used in this setting to clear HCV and, in doing so, to have a beneficial effect on renal injury. We hypothesized that we could obtain the same benefit for HCV-infected LT recipients treated with antiviral therapy. Our results appear to confirm previous results in immunocompetent patients, that is, an improvement in GFR in patients who cleared the virus but only when they had mild impairment of renal function (stage 2 CKD) before they started antiviral therapy. We had only a small number of patients with more advanced RD before treatment (n = 31), and we had no information on proteinuria at different time points for our patients or histological information on the type of kidney lesion. Although we checked whether changes in CNIs (drugs and trough levels) after antiviral therapy could have caused the amelioration in kidney function, we could not find significant differences in CNI trough levels and/or CNI discontinuation rates between SVR patients and NRs. A complete assessment of the overall potentially nephrotoxic immunosuppression that each patient receives over time is, however, a challenging task, and it is probably not appropriately assessed in most studies. Whether additional unmeasured measures or management modifications for SVR patients may have contributed to the improvement in renal function is unknown at the present time, but on the basis of an analysis of other factors potentially involved in renal function, we believe that some of the difference in kidney function at 5 years can be attributed to differences in rates of end-stage liver disease within the first 5 years after treatment. Indeed, the proportions of patients with stage 2 CKD before treatment (MDRD-4 value = 60-89 mL/minute) and with cirrhosis at the last follow-up were 22% for SVR patients and 56% for NRs. This could also explain why it takes 5 years of follow-up to see a significant difference in GFRs because the improvement in fibrosis typically reported for responders is known to take years to occur.
In stage 2 CKD patients, the median changes in the MDRD-4 values from the pretreatment period to 1, 3, and 5 years after treatment were −0.5, 4.5, and 9.4 mL/minute/1.73 m2, respectively, for SVR patients and −1, −0.3, and −1.5 mL/minute/1.73 m2, respectively, for NRs. One might question whether the difference in the final GFR values represents a clinically significant difference. More than the static difference between 9.4 and −1.5 mL/minute/1.73 m2 (which is statistically significant), we think that it is the trend observed over time that is clinically relevant. In addition, in several recent studies aimed at evaluating the potential benefits of new nonnephrotoxic immunosuppressive regimens after LT, a difference of 7 mL/minute between immunosuppressive arms was considered relevant.
All data in this study were analyzed according to when patients initiated their HCV treatment after LT. This created the potential for a lead-time bias because GFR generally declined over time in these patients and the patients may not have started HCV treatment at the same time after transplantation. In order to exclude this possibility, we plotted MDRD values versus the time after LT (months) as a function of the viral response to therapy. Before the initiation of therapy, similar trends in renal function were observed for the SVR and NR groups, with slightly worse MDRD values observed for those who subsequently achieved an SVR versus NRs. In contrast, during and after therapy, SVR patients had better MDRD values than NRs. Furthermore, no difference in the time from LT to therapy was observed for stage 2 patients.
The eradication of HCV has been shown in previous studies to reduce HCV-related fibrosis progression, to ameliorate HCV-related graft lesions, and, in doing so, to reduce the rate of hepatic decompensation and improve graft and patient survival.[1, 2, 4-6] We have shown in this study another benefit of viral clearance that may also have an impact on long-term patient survival.
The independent association between pretreatment RD and CKD at the last follow-up is logical and similar to the association found in several studies between pre-LT renal function and posttransplant CKD.
In conclusion, CKD is a common finding in LT recipients treated for HCV. In patients with mild CKD, viral eradication is associated with an improvement in renal function over time.