LIVER TRANSPLANTATION 16:440-446, 2010
ORIGINAL ARTICLE
Pretransplant Predictors of Recovery of Renal Function After Liver Transplantation Patrick G. Northup,1 Curtis K. Argo,1 Mihir R. Bakhru,1 Timothy M. Schmitt,2 Carl L. Berg,1 and Mitchell H. Rosner3 1 Division of Gastroenterology and Hepatology, 2Department of Surgery, and 3Division of Nephrology, University of Virginia, Charlottesville, VA
The Model for End-Stage Liver Disease system has given priority on the liver transplant waiting list to candidates with renal failure. This study determined the predictors of spontaneous recovery of renal function after transplantation in 1041 liver transplant recipients on renal replacement therapy (RRT) at the time of transplant (from February 2002 to January 2007). Data from these patients were obtained from the US Organ Procurement and Transplantation Network and US Renal Data System databases. Univariate and multivariate survival models were constructed along with multivariate logistic regression models to find independent predictors of spontaneous renal recovery. Seven hundred seven recipients (67.9%) had spontaneous recovery of renal function after liver transplantation. Those recovering spontaneously had a significantly shorter course of RRT in the pretransplant time period (15.6 versus 36.6 days, P < 0.001). Recovery of renal function was observed in 70.8% and 11.5% of recipients on RRT for less than 30 days and more than 90 days, respectively. Other statistically significant pretransplant variables independently associated with recovery of renal function included recipient age, recipient pretransplant diabetes, and donor age. In conclusion, the duration of pretransplant RRT is highly predictive of spontaneous renal recovery post-transplant. Liver transplant candidates requiring less than 30 days of pretransplant RRT are likely to spontaneously recover renal function after liver transplantation, whereas those on RRT for more than 90 days are not. Liver Transpl 16:440–446, 2010. V 2010 AASLD. C
Received August 24, 2009; accepted December 13, 2009. In the era of Model for End-Stage Liver Disease (MELD)–based liver allocation, there has been a preference for organ allocation to those patients with renal dysfunction, with the result that more than 30% of patients awaiting liver transplantation have decreased renal function, including a substantial number that require renal replacement therapy (RRT).1,2 This has coincided with a rise in combined liver-kidney transplantation (CLKT) in the past 7 years. Impaired renal
function at the time of transplantation has been shown to have a detrimental impact on morbidity and mortality after liver transplantation.2,3 Despite imprecise methods for measuring renal dysfunction,4 it is clear that the pretransplant degree of renal insufficiency has a major impact on post-transplant survival.5 Specifically, the length of pretransplant renal dysfunction predicts post-transplant renal insufficiency. Campbell et al.6 demonstrated that the duration of pretransplant
Abbreviations: CI, confidence interval; CLKT, combined liver-kidney transplantation; INR, international normalized ratio; MELD, Model for End-Stage Liver Disease; N/A, not available; OR, odds ratio; RRT, renal replacement therapy; UNOS, United Network for Organ Sharing; USRDS, US Renal Data System. Patrick G. Northup assisted with all aspects of this article. Curtis K. Argo assisted with the conception and design, the analysis and interpretation of the data, and a critical revision of the manuscript. Mihir R. Bakhru assisted with the analysis and interpretation of the data and the drafting of the manuscript. Timothy M. Schmitt assisted with the analysis and interpretation of the data and a critical revision of the manuscript. Carl L. Berg assisted with the conception and design, the analysis and interpretation of the data, the acquisition of the data, and a critical revision of the manuscript. Mitchell H. Rosner assisted with the conception and design, the analysis and interpretation of the data, the acquisition of the data, the acquisition of funding, and a critical revision of the manuscript. The principal author had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. The data reported here have been supplied by the US Renal Data System and the United Network for Organ Sharing. The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy or interpretation of the US Government. This work was partially funded by the Jan Albrecht Commitment to Clinical Research in Liver Disease Award from the American Association for the Study of Liver Diseases. Address reprint requests to Patrick G. Northup, M.D., M.H.E.S., Division of Gastroenterology and Hepatology, University of Virginia, Jefferson Park Avenue and Lee Street, MSB 2142, Charlottesville, VA 22908-0708; Telephone: 434-243-2718; FAX: 434-244-7529; E-mail:
[email protected] DOI 10.1002/lt.22008 Published online in Wiley InterScience (www.interscience.wiley.com).
C 2010 American Association for the Study of Liver Diseases. V
RENAL REPLACEMENT THERAPY AND LIVER TRANSPLANTATION 441
renal insufficiency was predictive of 6- and 12-month serum creatinine levels after liver transplantation.6 This study was confirmed by Bahirwani et al.,7 who reported similar results and found that renal dysfunction present for more than 12 weeks prior to transplant was a predictor of a decreased glomerular filtration rate after liver transplantation. For those with the most severe renal dysfunction, the duration of RRT has been shown to predict the post-transplant outcome. In a single-center study, Ruiz et al.8 assessed the survival of patients with pretransplant hepatorenal syndrome and demonstrated increased survival after CLKT in patients receiving pretransplant dialysis for more than 8 weeks. Some have recommended that patients receiving more than 6 weeks of RRT deserve consideration for CLKT.9 A review of Scientific Registry of Transplant Recipients data during a recent consensus meeting suggested that few patients with pretransplant renal dysfunction are listed for subsequent kidney transplantation 1 year after orthotopic liver transplantation, and this indicates that most of these patients spontaneously recover renal function or that renal function stabilizes.9 This may be especially true for those patients with hepatorenal syndrome. Because of this, policymakers in the United States are considering formalizing criteria for simultaneous liver-kidney transplantation.10 Nonetheless, controversy exists regarding which patients derive benefit from combined transplantation and which patients have reversible kidney injury that will eventually recover. It is often difficult to define with certainty the cause of renal dysfunction in these critically ill patients, and this further makes prognostic decision making difficult in this population. Given the shortage of available organs, the need for CLKT must be better defined in order to avoid unnecessary renal allocation to those who will have spontaneous improvement of renal function. The aim of our study was to determine the outcomes and characteristics of patients on RRT at the time of liver transplantation that predict eventual recovery of renal function after isolated liver transplantation.
PATIENTS AND METHODS Pretransplant candidate information was obtained from the United Network for Organ Sharing (UNOS)/ Organ Procurement and Transplantation Network database (http://www.unos.org). This de-identified data set contains detailed information on candidates’ pretransplantation phase, including the length of RRT, laboratory values, ascites, encephalopathy levels, and MELD components. Validated post-transplant survival, donor information, and graft status are also incorporated into the UNOS data set. All adults (18 years old and older) who underwent liver transplantation between February 27, 2002 and January 18, 2007 were analyzed for inclusion in the study. In order to exclude the possibility of calcineurin inhibitor toxicity as a confounder, all liver retransplants were excluded. Multivisceral transplants and status 1 (acute liver failure) transplants were also excluded.
No information on pretransplant renal disease etiology is included in the UNOS data set. In order to assess the spontaneous recovery of function of recipients’ native kidneys, patients that underwent CLKT were excluded as well. All other patients needing either continuous or intermittent (at least twice weekly) RRT prior to transplantation were included in the analysis. The UNOS data set does not contain dependable information on post-transplantation renal function. To obtain long-term post-transplant follow-up information, the US Renal Data System (USRDS) data set (http://www.usrds.org) was analyzed in conjunction with data from the aforementioned UNOS subset. This data set contains detailed information on all US patients requiring RRT for more than 30 contiguous days and captures more than 97% of this population.11 With a custom data set merge, transplant recipients listed in the UNOS data set were matched with corresponding entries in the USRDS data set, and renal outcome data for these recipients were merged into the final study data set. Recipients were characterized as either spontaneously recovering renal function after transplantation or not spontaneously recovering. Recipients who underwent transplantation while on RRT and survived more than 30 days after liver transplant but did not have a corresponding entry in the USRDS data set were assumed to have spontaneously recovered renal function in less than 30 days. Those patients that eventually received a renal transplant after liver transplantation were not considered to have spontaneously recovered renal function, although a separate survival analysis was performed for this group (see the Results section). Because the glomerular filtration rate and details of renal function are not reported in the USRDS data set at the time of removal from RRT, spontaneous renal function recovery was defined as removal from RRT without death or renal transplantation. Study subjects were characterized by demographics, pretransplant liver function, the presence of pretransplant diabetes mellitus, and the duration of pretransplant RRT in 30-day increments. Transplant and donor characteristics were also assessed, including the components of the donor risk index,12 terminal laboratory values of the donor, and diabetes mellitus status. Differences in categorical variables between recipients spontaneously recovering renal function and those that did not were compared with the chisquare test or Fisher exact test as appropriate. Continuous variables were compared with the Student t test or Wilcoxon signed-rank test as appropriate. Multivariate logistic regression models were constructed to assess for adjusted independent variables associated with the spontaneous recovery of renal function. Unadjusted post-transplant survival was estimated by the Kaplan-Meier technique, and differences were measured by the log-rank test. A multivariate survival model was constructed with Cox proportional hazards techniques, and differences were compared by an analysis of maximum likelihood estimates. Candidate variables were included in the
LIVER TRANSPLANTATION.DOI 10.1002/lt. Published on behalf of the American Association for the Study of Liver Diseases
442 NORTHUP ET AL.
TABLE 1. Characteristics of Patients Undergoing Liver Transplantation Who Were on RRT for at Least One Week Before Transplantation
Recipient factors Age, years, mean (95% CI) Male, n (%) African American, n (%) MELD score at transplant, mean (95% CI) Serum bilirubin at transplant, mg/dL, mean (95% CI) Serum INR at transplant, mean (95% CI) Serum creatinine at registration, mg/dL, mean (95% CI) Serum albumin at transplant, g/dL, mean (95% CI) Days of pretransplant RRT, mean (95% CI) Hepatitis C infection, n (%) Severe ascites prior to transplant, n (%)* Severe encephalopathy prior to transplant, n (%)* Diabetes mellitus, n (%) Body mass index at registration, kg/m2, mean (95% CI) Donor and procedural factors Age, years, mean (95% CI) Terminal creatinine, mg/dL, mean (95% CI) Diabetes mellitus, n (%) Waiting list time, days, mean (95% CI) Cold ischemia time, hours, mean (95% CI) Donor risk index, mean (95% CI)y Length of stay after transplant, days, mean (95% CI) Total bilirubin on discharge, mg/dL, mean (95% CI)
No Spontaneous
Spontaneous
Recovery of Renal Function
Recovery of Renal Function
(n ¼ 334)
(n ¼ 707)
P Value
53.2 (52.2-54.2) 224 (67.1) 31 (9.3) 34.5 (33.5-35.5) 17.1 (15.3-18.9) 2.30 (1.91-2.69) 2.98 (2.71-3.25) 2.89 (2.82-2.97) 36.6 (26.0-47.1) 116 (34.7) 183 (54.8) 101 (30.2) 106 (31.7) 28.9 (28.1-29.6)
51.9 (51.3-52.6) 478 (67.6) 44 (6.2) 36.7 (36.1-37.3) 19.1 (18.0-20.3) 2.38 (2.28-2.48) 2.39 (2.24-2.53) 2.89 (2.84-2.94) 15.6 (11.1-20.0) 290 (41.0) 415 (58.7) 224 (31.7) 139 (19.7) 30.6 (27.2-34.0)
0.04 0.86 0.07
