Chronic kidney disease after liver transplantation in human immunodeficiency virus/hepatitis C virus–coinfected recipients versus human immunodeficiency virus–infected recipients without hepatitis C virus: Results from the national institutes of health multi-site study
Division of Gastroenterology and Hepatology, University of Pennsylvania, Philadelphia, PA
This study was supported by the Solid Organ Transplantation in HIV: Multi-Site Study (AI052748), which is funded by the National Institute of Allergy and Infectious Diseases (NCT00074386 at ClinicalTrials.gov). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Allergy and Infectious Diseases or the National Institutes of Health.
Address reprint requests to K. Rajender Reddy, M.D., Division of Gastroenterology and Hepatology, University of Pennsylvania, 2 Dulles, 3400 Spruce Street, Philadelphia, PA 19104. Telephone: 215-662-4311; FAX: 215-615-1601; E-mail: email@example.com
Chronic kidney disease (CKD), defined by the National Kidney Foundation as the presence of a reduced glomerular filtration rate for 3 months or longer, is a common complication after orthotopic liver transplantation (OLT) and has a major impact on graft and patient survival.[1, 2] The presence of CKD after transplantation not only is important with respect to the need for renal replacement therapy but also increases the risk for cardiovascular disease and mortality; this emphasizes the need for protection of renal function after liver transplantation. Hepatitis C virus (HCV) infection has been shown to be a risk factor for CKD in liver transplant recipients because of an increased incidence of glomerular disease as well as the association between HCV and diabetes, which is an established CKD risk factor.[4-6]
It is estimated that approximately 25% to 30% of patients with human immunodeficiency virus (HIV) are coinfected with HCV. Individuals coinfected with both viruses are at risk for accelerated liver disease and, consequently, cirrhosis, liver failure, and hepatocellular carcinoma. HIV infection has also been associated with a spectrum of kidney diseases, including HIV-associated nephropathy (a major cause of nephrotic syndrome in HIV-infected patients), immune complex glomerulonephritis, thrombotic microangiopathy, and nephrotoxicity from antiretroviral therapies such as the nucleotide reverse transcriptase inhibitor tenofovir (proximal tubular dysfunction) and the protease inhibitor (PI) indinavir (nephrolithiasis, obstructive nephropathy, and interstitial nephritis).[9, 10]
Because an increasing number of patients coinfected with HIV and HCV are undergoing liver transplantation, the impact of HIV/HCV coinfection on posttransplant CKD is an important outcome to study, particularly in the setting of calcineurin inhibitor immunosuppression. The aim of this study was to evaluate the incidence of posttransplant CKD in HIV/HCV-coinfected liver transplant recipients with the hypothesis that these patients would have worse posttransplant kidney function than HIV-infected transplant recipients without HCV (HIV/non-HCV recipients).
PATIENTS AND METHODS
Data obtained from a prospective, multicenter study of 125 HIV-infected OLT recipients from 2003 to 2010 (the National Institutes of Health Solid Organ Transplantation in HIV: Multi-Site Study) were analyzed in order to assess the pretransplant prevalence of CKD [estimated glomerular filtration rate (eGFR) < 60 mL/minute for ≥3 months] and the incidence of CKD 1 and 3 years after transplantation. Institutional review board approval was obtained prior to study initiation. There were 9 simultaneous liver/kidney transplant recipients in the database, and these patients were excluded from our analysis. The remaining 116 patients were divided into 2 cohorts: patients coinfected with HIV and HCV and patients without HCV (Fig. 1). Eighty-one patients in the cohort were HIV/HCV-coinfected, and 35 patients were not infected with HCV [23 were hepatitis B virus (HBV)–infected]. Although data on matched HCV-monoinfected patients were collected in the prospective database, only pretransplant kidney function data were available, so posttransplant renal outcomes could not be assessed in this cohort as a comparison group. Data from the Scientific Registry of Transplant Recipients (SRTR) were analyzed to assess the prevalence of pretransplant CKD and the incidence of CKD after transplantation in all OLT-alone recipients with a positive HCV serology during a similar time frame.
The primary outcome that was assessed was the incidence of CKD 1 and 3 years after transplantation, with the focus being a comparison of HIV/HCV-coinfected patients and HIV-positive patients without concomitant HCV. The eGFR was determined with the 4-point Modification of Diet in Renal Disease equation, which is based on serum creatinine, age, sex, and race. Baseline and quarterly eGFR levels through year 3 after transplantation were used in the analysis.
Variables analyzed in our study included age (donor and recipient), race, sex, HIV viral load and CD4 count (both before and after transplantation), eGFR at transplant, highly active antiretroviral therapy (HAART) regimen [defined as time-varying covariates of tenofovir, stavudine (D4T)/didanosine (DDI), or any PI use within the previous 3 months], presence of diabetes (before transplantation and de novo after transplantation; implied in the database by the use of insulin or oral hypoglycemic therapy), hypertension (defined by the use of antihypertensives), most recent Model for End-Stage Liver Disease (MELD) score before transplantation, incidence of acute cellular rejection, cytomegalovirus (CMV) infection, HCV therapy (after transplantation), presence of proteinuria after transplantation, HBV status, serum albumin, initial immunosuppression regimen (tacrolimus versus cyclosporine), and cyclosporine use in previous 3 months (as a time-varying covariate).
Statistical analyses were performed with SAS 9.2 (SAS Institute, Cary, NC). A 2-sided P value less than 0.05 was considered to indicate statistical significance. Descriptive statistics included percentages, medians, and interquartile ranges (IQRs) as appropriate. A comparison of baseline characteristics was conducted with Fisher's exact test (categorical variables) or the Wilcoxon rank-sum test (continuous variables).
A Kaplan-Meier analysis was performed for the time to the development of CKD after transplantation in patients without pretransplant renal dysfunction. Patients were censored at the last follow-up. Estimated rates of CKD were calculated with the Kaplan-Meier method, and 95% confidence intervals (CIs) were estimated with Greenwood's formula. Survival distributions were compared with the log-rank test. For subjects without pretransplant CKD, proportional hazards models were performed to assess predictors of posttransplant CKD and stage 4/5 CKD. All variables with P < 0.1 from the univariate model were included in an initial multivariate model. Subsequently, variables with P ≥ 0.1 were excluded, the model was refit, and all interactions were examined. The impact of posttransplant CKD (as a time-varying covariate) on graft loss and death was also evaluated in our cohort via univariate proportional hazards models.
Seventy-eight percent of the 116 study patients were male, 69% were Caucasian, and the median age at transplant was 48 years (IQR = 43-52 years). The median pretransplant serum creatinine level was 1.1 mg/dL (IQR = 0.8-4.4 mg/dL), and the median eGFR was 77 mL/minute (IQR = 56-111 mL/minute). The median posttransplant follow-up in this study was 3.6 years (IQR = 2.2-5.0 years).
Thirty-four of the 116 patients in the study cohort (29%) were suspected to have pretransplant CKD with an eGFR < 60 mL/minute; 25 of these patients were HIV/HCV-coinfected, and 9 were HIV-infected but not HCV-infected. Twenty-two patients had an eGFR between 30 and 59 mL/minute (stage 3 CKD), 9 between 15 and 29 mL/minute (stage 4 CKD), and 3 patients had an eGFR between 10 and 14 mL/minute (stage 5 CKD) before transplantation. Twenty of these 34 patients (59%) had at least stage 3 CKD after transplantation (6 had stage 4/5 CKD). In the cohort assessed with SRTR data, 33% of the 11,436 liver transplant recipients identified with a positive HCV serology were suspected to have pretransplant CKD (22% had stage 3 CKD, 8% had stage 4 CKD, and 3% had stage 5 CKD).
Twenty-six of the 82 patients without pretransplant CKD were HIV/non-HCV, and 56 were HIV/HCV-coinfected. The median age of this cohort was 48 years (IQR = 43-52 years). Among the 6594 SRTR HCV/non-HIV patients without pretransplant CKD, the median age at transplant was 54 years (IQR = 50-57 years), which was significantly higher than the median age for the HIV cohort (P < 0.001). The median MELD score at transplant was significantly lower for the HCV/non-HIV group (15) versus the HIV/HCV-coinfected group (18) and the HIV/non-HCV group (16; P = 0.005). There were no differences between the applicable groups with respect to pretransplant characteristics, including eGFR at transplant (P = 0.59), pre-OLT diabetes (P = 0.64), sex (P = 0.54), and race (P = 0.24; Table 1). HIV RNA was undetectable in 88% of the HIV/HCV-coinfected patients at enrollment versus 85% in the HIV-monoinfected group (P = 0.74). The median donor age was 43 years in the HIV/HCV-coinfected and HCV-monoinfected groups and 42 years in the HIV-monoinfected group (P = 0.80).
Table 1. Characteristics of OLT Recipients Without CKD Before Transplantation
HIV/HCV Patients (n = 56)
HIV/Non-HCV Patients (n = 26)
HCV/Non-HIV Patients (n = 6594)
The data are presented as medians and IQRs.
Based on the use of insulin or any glucose-lowering therapy.
Based on available laboratory data at select time points and posttransplant adverse events.
Posttransplant characteristics, including the incidence of acute cellular rejection (P = 0.63), CMV infection (P = 0.30), diabetes (P = 0.34), and hypertension (P = 0.31), also were not significantly different between the HIV/HCV-coinfected and HIV/non-HCV groups (Table 1).
Forty-one of the 82 patients without pretransplant CKD (50%) experienced CKD after transplantation; 11 of these patients developed stage 4/5 CKD (5 within 3 years and 6 between 3 and 5 years after transplantation). Information on the need for renal replacement therapy was not systematically collected in the database, so the incidence of end-stage renal disease could not be determined. The overall cumulative rates of stage 3 CKD and stage 4/5 CKD in the HIV cohort were 30% (95% CI = 21%-43%) and 1% (95% CI = 0%-9%), respectively, 1 year after transplantation. Three years after transplantation, the corresponding cumulative rates of stage 3 CKD and stage 4/5 CKD were 62% (55% in HIV/non-HCV patients versus 65% in HIV/HCV-coinfected patients) and 8% (0% in HIV/non-HCV patients versus 12% in HIV/HCV-coinfected patients), respectively (Table 2). Eight of the 15 HIV/HBV-coinfected patients without pretransplant CKD in our cohort developed stage 3 CKD after transplantation, and none of these patients developed stage 4/5 CKD. For the SRTR HCV/non-HIV subjects without pretransplant CKD, the cumulative rates of stage 3 CKD and stage 4/5 CKD were 23% and 2.2%, respectively, 1 year after transplantation. Three years after transplantation, the corresponding cumulative rates of stage 3 CKD and stage 4/5 CKD were 32% and 6.1%. The times to the development of CKD and stage 4/5 CKD are shown in Kaplan-Meier plots (Figs. 2 and 3). The cumulative incidence of CKD was significantly lower for HCV/non-HIV controls versus HIV/HCV-coinfected and HIV/non-HCV subjects (log-rank test, P < 0.001). On the other hand, the cumulative incidence of stage 4/5 CKD was significantly higher for HIV/HCV-coinfected subjects versus both HIV/non-HCV and HCV/non-HIV subjects (log-rank test, P = 0.001).
Table 2. Cumulative Incidence of CKD After OLT in Patients Without CKD Before Transplantation
Entire HIV Cohort
HIV/HCV Patients (n = 56)
HIV/Non-HCV Patients (n = 26)
SRTR HCV/Non-HIV Patients (n = 6594)
NOTE: The data are presented as percentages; 95% CIs are shown in parentheses.
In the multivariate proportional hazards model, a higher CD4 count (as a time-varying covariate) after transplantation was associated with a significantly lower risk of CKD [hazard ratio (HR) = 0.90 per 50 cells/μL, P = 0.01], and older age was associated with a greater incidence of CKD (HR = 1.05 per year, P = 0.03). A higher baseline eGFR was marginally associated with a lower incidence of CKD (HR = 0.99 per mL/minute, P = 0.08). Cyclosporine use in the previous 3 months, another significant predictor in the univariate model (HR = 1.87, P = 0.048), lost its significance in the multivariate model. HBV coinfection, HCV coinfection, race, diabetes, hypertension, proteinuria, and the HAART regimen had no significant impact on the development of CKD after transplantation (Table 3).
Table 3. Univariate/Multivariate Proportional Hazards Regression Models for CKD
There were only 11 cases of stage 4/5 CKD after transplantation in our study. In the univariate proportional hazards regression models, HCV coinfection (HR = 9.74, 95% CI = 1.22-77.8, P = 0.03), age (HR = 1.09, 95% CI = 1.01-1.18, P = 0.02), and female sex (HR = 3.94, 95% CI = 1.19-13.1, P = 0.03) were the only factors significantly associated with stage 4/5 CKD. In the multivariate model, HCV coinfection (HR = 10.8, 95% CI = 1.30-89.4, P = 0.03) and age (HR = 1.12, 95% CI = 1.01-1.25, P = 0.03) remained statistically significant, whereas female sex was excluded (P = 0.09; Table 4).
In the univariate proportional hazards models, posttransplant CKD was not significantly associated with an increased risk of death (HR = 1.24, 95% CI = 0.51-3.00, P = 0.64) or graft loss (HR = 0.87, 95% CI = 0.37-2.02, P = 0.74) after transplantation.
The administration of effective antiretroviral therapy has changed the natural course of HIV and led to dramatic improvements in patient survival. As a result, liver disease is now the leading cause of mortality among HIV-positive individuals and particularly among those coinfected with chronic HCV, and HIV is no longer considered a contraindication to OLT.[13, 14] CKD is a major contributor to post–liver transplant morbidity and mortality, and both HIV and HCV have been associated with kidney diseases of several origins. There have been limited data thus far on renal outcomes after liver transplantation in patients coinfected with HIV and HCV.
The results of our study reveal that CKD is prevalent both before and after transplantation in HIV-positive liver transplant recipients. Among the 116 patients evaluated in the study, 29% (34 patients) had pretransplant CKD, and 12 patients (approximately 10% of the cohort) had stage 4/5 CKD before transplantation. Despite national guidelines that suggest the consideration of simultaneous liver-kidney transplantation in patients with stage 4 CKD, these patients underwent only liver transplantation. Importantly, only 59% of the patients suspected to have pretransplant CKD had an eGFR < 60 mL/minute after transplantation. Six patients (18% of the cohort with pretransplant CKD) had stage 4/5 CKD after transplantation. Because pretransplant renal dysfunction is the most important determinant of posttransplant CKD, the low rates of posttransplant CKD in this subset of patients are reassuring. This may potentially reflect the emphasis on renal-sparing immunosuppression strategies in this patient cohort; however, these optimistic results could also be due to the limited median posttransplant follow-up of 3.6 years.
Among the patients without baseline CKD before transplantation, 62% developed stage 3 CKD and 8% developed stage 4/5 CKD within 3 years after transplantation. In the univariate analysis, pretransplant eGFR, a known predictor of posttransplant renal function, was significantly associated with posttransplant CKD. However, in the multivariate analysis, eGFR was marginally predictive of posttransplant CKD.
Despite our hypothesis, HCV coinfection was not a significant predictor of posttransplant CKD in HIV-positive liver transplant recipients in our study in univariate or multivariate analyses. The lack of an association between HCV and posttransplant CKD may be attributable to the similar rates of diabetes in the HIV/HCV-coinfected and HIV/non-HCV cohorts in our study. Diabetes has been cited as a potential mechanism for worse renal outcomes in HCV-infected liver transplant recipients. Furthermore, recent data have revealed no worse posttransplant renal outcomes for patients with HCV cirrhosis and normal pre–liver transplant renal function in comparison with patients with other causes of liver disease. Using SRTR data, an analysis of liver transplant recipients with a positive HCV serology during a similar time frame also revealed a lower cumulative incidence of stage 3 CKD (32%) and stage 4/5 CKD (6.1%) 3 years after transplantation in comparison with the HIV/HCV-coinfected cohort, whose cumulative rates of stage 3 CKD and stage 4/5 CKD were 65% and 12%, respectively.
Although there was no association between HCV and posttransplant CKD in our study, all patients with normal pretransplant renal function who developed at least stage 4 CKD after transplantation were HIV/HCV-coinfected. Furthermore, the cumulative incidence of stage 4/5 CKD was significantly higher in the HIV/HCV-coinfected cohort versus the HIV/non-HCV cohort (12% versus 0%; Fig. 3). The event numbers were small for a robust analysis, but this raises the possibility of an association that could be confirmed with a larger sample size and/or a longer duration of follow-up because our multivariate analysis did suggest an association between HCV coinfection and stage 4/5 CKD after transplantation (HR = 10.8, P = 0.03). Our results suggest a possible synergistic effect of HIV and HCV coinfection on the development of advanced posttransplant renal disease because of the higher incidence of stage 4/5 CKD in these patients versus HIV/non-HCV and HCV-monoinfected controls.
Older age and lower CD4 counts were significant predictors of posttransplant CKD in our study in the multivariate analysis. Increasing age is a well-known predictor of posttransplant CKD with a relative risk per 10-year increment of 1.36 according to observations by Ojo et al. in a population-based cohort study assessing the risk factors for and incidence of CKD in nonrenal transplant recipients. In our study, there was a 5% increase in the risk of CKD for every year increment in the recipient's age (P = 0.03). Higher CD4 counts were protective against posttransplant CKD in our study. Perhaps this was due to a reduced incidence of HIV nephropathy after transplantation, which is a common cause of CKD in HIV-infected individuals associated with a low CD4 count. Unfortunately, we do not have comprehensive data on posttransplant proteinuria or histological data from renal biopsy samples to assess the incidence of HIV-associated focal segmental glomerulosclerosis in this cohort.
HAART therapy was not associated with posttransplant CKD in our analysis. Because of the significant drug interactions between antiretroviral therapy and calcineurin inhibitor–based immunosuppression, it is possible that vigilant monitoring of renal function in these patients resulted in fewer cases of posttransplant CKD, so no association was found between HAART and renal dysfunction in univariate and multivariate analyses. Cyclosporine use in the previous 3 months, which was a significant predictor in the univariate model (HR = 1.87, P = 0.048), lost its significance in the multivariate model for CKD. The impact of immunosuppression trough levels was also assessed, but no association with posttransplant CKD was observed.
We also assessed the impact of posttransplant CKD on graft loss and death and found no significant association between posttransplant renal dysfunction and graft loss (P = 0.74) or mortality (P = 0.64). This was likely due to the small percentage of patients with significant posttransplant CKD and the relatively abbreviated follow-up of our study. A recent publication from the National Institutes of Health–sponsored Solid Organ Transplantation in HIV: Multi-Site Study revealed 3-year patient and graft survival rates of 60% and 53%, respectively, in HIV/HCV-coinfected patients.
The strength of our study lies in its large sample size. The National Institutes of Health Solid Organ Transplantation in HIV: Multi-Site Study represents the largest cohort of HIV-positive transplant recipients from 21 centers in the United States. Its limitations include a median follow-up of approximately 3.6 years after transplantation, which makes it difficult to assess longer term renal outcomes after liver transplantation. Additionally, we did not have thorough data on pretransplant and posttransplant proteinuria in our patients, which is an important variable for the assessment of kidney disease. Furthermore, histological data regarding the etiology of posttransplant kidney disease were not assessed in our study in order to determine the mechanism of posttransplant CKD in our cohort. Another limitation of our analysis is the lack of HCV RNA status data for the SRTR HCV-monoinfected control group, so it is possible that some of the controls did not have HCV viremia at the time of transplantation. Nonetheless, our data reveal that CKD is prevalent both before and after liver transplantation in HIV/HCV-coinfected patients, and the incidence of advanced CKD was 12% (95% CI = 5%-28%) 3 years after transplantation in this patient cohort. Important predictors of CKD include increasing age and lower CD4 counts after transplantation. HCV coinfection may be associated with a higher incidence and burden of advanced CKD in the long term.
The authors thank all of the Solid Organ Transplantation in HIV: Multi-Site Study investigators and coordinators for their hard work and dedication to the study and subjects.