The decision to perform combined kidney/liver

Transcription

1 ORIGINAL ARTICLES Renal Function after Orthotopic Liver Transplantation is Predicted by Duration of Pretransplantation Creatinine Elevation Mical S. Campbell, 1 David S. Kotlyar, 2 Colleen M. Brensinger, 3 James D. Lewis, 1,3 Kirti Shetty, 4 Roy D. Bloom, 5 James F. Markmann, 6 Kim M. Olthoff, 6 Abraham Shaked, 6 and K. Rajender Reddy 1 See Editorial on Page 1022 Abbreviations: CKLT, combined kidney/liver transplantation; OLTa, orthotopic liver transplantation alone; OLT, orthotopic liver transplantation; MELD, model for end-stage liver disease; HRS, hepatorenal syndrome; RRT, renal replacement therapy; ROC, receiver operating characteristic. From the 1 Division of Gastroenterology, University of Pennsylvania Health System, Philadelphia, PA; the 2 University of Pennsylvania School of Medicine; the 3 Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania; the 4 Division of Gastroenterology, Georgetown University Hospital; the 5 Department of Nephrology, University of Pennsylvania Health System; and the 6 Division of Surgery, University of Pennsylvania Health System. Received February 25, 2005; accepted March 7, Address reprint requests to K. Rajender Reddy, MD, Hospital of the University of Pennsylvania, Division of Gastroenterology, 3 Ravdin, 3400 Spruce St., Philadelphia, PA Telephone: ; FAX: ; rajender.reddy@uphs.upenn.edu Copyright 2005 by the American Association for the Study of Liver Diseases Published online in Wiley InterScience ( DOI /lt In patients with recent onset renal insufficiency, the decision to perform combined kidney/liver transplantation (CKLT) vs. orthotopic liver transplantation alone (OLTa) can be difficult. We hypothesized that duration of renal dysfunction may correlate with creatinine elevation after liver transplantation. We retrospectively identified 69 liver transplantation patients with pretransplantation creatinine >1.5 mg/dl (53 OLTa, 13 CKLT). Variables analyzed were presence of hepatorenal syndrome, creatinine, Model for End-Stage Liver Disease score, albumin, age, race, gender, cause of liver disease, diabetes mellitus, hypertension, and history of ascites, spontaneous bacterial peritonitis, variceal bleeding, hepatic encephalopathy, renal replacement therapy (RRT), and transjugular intrahepatic portosystemic shunting. Duration of pretransplantation renal dysfunction was predictive of 6- and 12-month creatinine post-olta. Area under the receiver operating characteristic (ROC) curve for prediction of 12-month renal insufficiency by renal dysfunction duration was 0.71; optimal duration cutoff was 3.6 weeks. We applied a multivariable model, derived from OLTa patients, to CKLT subjects with definite or possible hepatorenal syndrome. Predicted 12-month creatinine without renal transplantation was >2.0 mg/dl for each patient. CKLT patients as opposed to OLTa patients had longer duration of renal dysfunction (median, 18.1 vs. 2.7 weeks, P < 0.001), higher creatinine (median 4.0 versus 1.7 mg/dl, P < 0.001), and higher rate of pretransplantation RRT (62% vs. 7%, P < 0.001). Adjusting for baseline characteristics, CKLT patients had lower creatinine than OLTa patients at 6 months (P 0.15) and 12 months (P 0.01) after transplantation. In conclusion, duration, but not cause, of renal dysfunction predicts renal outcome in OLTa recipients. Prospective studies may use duration of renal dysfunction to help identify CKLT candidates. (Liver Transpl 2005;11: ) The decision to perform combined kidney/liver transplantation (CKLT) as opposed to orthotopic liver transplantation alone (OLTa) can be difficult in patients with end-stage liver disease and recent onset renal insufficiency. The rate of acute renal failure among patients awaiting orthotopic liver transplantation (OLT) and the waiting time for OLT have increased in recent years. 1 The recent introduction of the Model for End-Stage Liver Disease (MELD) score will likely further enrich the proportion of OLT candidates who have renal dysfunction, as creatinine is a key component of MELD calculation. Because of scarce organ resources, it is important to predict accurately which patients with pretransplantation renal dysfunction will recover after OLT and who will have persistent or progressive kidney disease. Pretransplantation serum creatinine level is an important predictor of post-olt survival and renal dysfunction. 2-7 Even relatively mild elevations in preoperative creatinine ( mg/dl) may portend poor renal function postoperatively. 5-7 Cause of renal disease may also help predict posttransplantation creatinine. Certainly patients with underlying chronic kidney diseases such as glomerulonephritis, diabetic nephropathy, and polycystic kidney disease would be expected to have persistently poor or worsening renal function after OLTa, particularly in the setting of cal Liver Transplantation, Vol 11, No 9 (September), 2005: pp

2 Duration of Renal Dysfunction and OLT 1049 cineurin inhibitor based immunosuppression. Many transplant centers have reported that a large majority of their CKLT patients underwent transplantation for chronic kidney disease In contrast, studies from the early 1990s demonstrate that patients with hepatorenal syndrome (HRS) have a good post-olta renal outcome 14,15 and may avoid concomitant renal transplantation. Because waiting times for liver transplantation and duration of renal dysfunction prior to transplantation have increased since then, 16 it is possible that renal outcomes after OLTa in patients with HRS may be less favorable now. Recommendations from a recent review emphasize that cause and severity of renal disease should guide the decision to perform CKLT as opposed to OLTa. Furthermore, the authors suggest that in the absence of parenchymal renal disease, patients with duration of renal dysfunction 12 weeks will likely recover renal function and should not undergo CKLT. 16 However, data to support the premise that duration of pretransplantation renal dysfunction is an important predictor of post-olt renal function are not available. We performed our study to determine the association between duration of preoperative creatinine elevation and posttransplantation renal function. Patients and Methods Study Design We identified a retrospective cohort of 69 patients who underwent OLT between March 2000 and August 2003 at the Hospital of the University of Pennsylvania with renal dysfunction. Only patients with 2 serum creatinine values 1.5 mg/dl were included in the study. The database is a prospectively maintained repository of all clinical information related to liver transplantation patients seen at our institution and is well-suited to rigorous investigations. Six months to 4 years of follow-up laboratory and clinical information are available for all patients included in the study. Subjects who had previously undergone liver transplantation or who had fulminant hepatic failure were not included. Thirteen of 69 patients underwent CKLT. Institutional Review Board approval was obtained before the study was conducted. Our institution is a large liver transplantation center, performing liver transplantations per year. Patients included in the study were transplanted before and after the introduction of MELD score (February 27, 2002). Patients with renal disease were evaluated by our center s transplant nephrologists. The decision to perform CKLT vs. OLTa was made on a case-by-case basis. Factors that influenced the recommendation for CKLT included cause, duration, and severity of renal disease, as well as requirement for renal replacement therapy (RRT). All serum creatinine values before transplantation were obtained, starting with the first creatinine 1.5 mg/dl. Duration of renal dysfunction is defined as the number of weeks before transplantation during which creatinine was persistently 1.5 mg/dl, including the last creatinine prior to transplantation. Cause of renal disease was agreed upon for each patient by 3 investigators (M.S.C., D.S.K., and K.R.R.) and was classified as HRS, not HRS, or uncertain. Data used for classification of renal disease, obtained from both inpatient and outpatient records, include urine sodium, urinalysis, 24-hour urine volume, presence or absence of diuretics, response to intravenous fluid challenge, renal biopsy (if available), and pattern of changes in creatinine. Underlying conditions associated with chronic kidney disease, such as diabetes mellitus and hypertension, as well as exposure to acute insults such as intravenous contrast agents, nephrotoxic drugs, diuretics, and sepsis were also reviewed. Diagnosis of HRS was guided by International Ascites Club recommendations: presence of advanced liver disease, serum creatinine 1.5 mg/dl, absence of concurrent explanations for renal insufficiency, failure of response to volume expansion, and absence of proteinuria or other evidence for parenchymal renal disease, with additional criteria of low urine volume, low urine sodium, low serum sodium, urine osmolality plasma osmolality, and absence of red blood cells on urinalysis. 17 Patients with HRS were not subdivided into types 1 and 2 because the principal distinction between the 2 groups is duration of renal dysfunction, a variable analyzed separately. Other variables include presence of diabetes mellitus, hypertension, and hepatocellular carcinoma, cause of liver disease, age, race, gender, requirement for RRT prior to OLT, last serum creatinine prior to transplantation, MELD score, serum albumin, international normalized ratio, bilirubin, and history of ascites, spontaneous bacterial peritonitis, variceal bleeding, hepatic encephalopathy, transjugular intrahepatic portosystemic shunt placement, and type of immunosuppression given post-olt. Laboratory values closest in time to transplantation were used. Primary outcomes were serum creatinine values at 6 and 12 months posttransplantation. Secondary outcomes were requirement for RRT at any time post-olt and continuing RRT requirement at 6 and 12 months post-olt. Statistical Methods Means and standard deviations are reported for parametric variables. For nonnormally distributed variables, medians and quartiles are given. Wilcoxon rank sum tests, ANOVA, Fisher exact tests, and unpaired t tests were used to compare variables between groups undergoing OLTa and CKLT. Univariate and multivariate linear regression were performed for OLTa patients, forcing duration of pretransplant renal disease into multivariate models. A multivariable model including all patients was generated, forcing the treatment group (CKLT vs. OLTa) into the model. Pre-OLT creatinine could not be included in this model because it was collinear with the treatment group. Nonparametric variables were first transformed

3 1050 Campbell et al. Table 1. Baseline Characteristics of OLTa and CKLT Patients Variable OLTa (n 53) Mean Standard Deviation, n (%), or Median (Interquartile Range) CKLT (n 13) Mean Standard Deviation, n (%), or Median (Interquartile Range) P Value Age years years 0.40 Gender 45 (80%) male 11 (85%) male 1.00 Race 50 (89%) white 9 (69%) white 0.08 Required RRT 4 (7%) 8 (62%) Diabetes mellitus 15 (27%) 6 (50%) 0.17 Hypertension 20 (36%) 5 (42%) 0.75 Cause of renal disease 0.83 HRS 15 (27%) 2 (15%) Not HRS 32 (57%) 9 (69%) Uncertain 9 (16%) 2 (15%) Cause of liver disease 0.09 Viral 28 (50%) 6 (46%) NASH 12 (21%) 0 (0%) Alcohol 8 (14%) 2 (15%) Other 8 (14%) 5 (38%) Hepatocellular carcinoma present 10 (18%) 4 (31%) 0.44 Ascites 55 (98%) 12 (92%) 0.34 Hepatic encephalopathy 46 (82%) 9 (69%) 0.44 History of spontaneous bacterial peritonitis 21 (38%) 3 (23%) 0.52 History of variceal bleeding 14 (25%) 0 (0%) 0.06 TIPS placement 7 (12.5%) 3 (23%) 0.38 MELD Albumin 2.5 mg/dl ( mg/dl) 2.3 mg/dl ( mg/dl) 0.82 Pretransplant creatinine 1.7 mg/dl ( mg/dl) 4.0 mg/dl ( mg/dl) INR 1.9 ( ) 1.5 ( ) 0.26 Bilirubin 3.1 mg/dl ( mg/dl) 1.4 mg/dl ( mg/dl) 0.06 Duration of pretransplant renal disease 2.7 weeks ( weeks) 18.1 weeks ( weeks) Abbreviations: NASH, nonalcoholic steatohepatitis; TIPS, transjugular intrahepatic portosystemic shunt; INR, international normalized ratio. (logarithmic transformation for duration of renal disease, international normalized ratio, and bilirubin; reciprocal transformation for creatinine) before linear regression was performed. Multivariable linear regression models were generated by stepwise backward elimination, using a P value greater than 0.05 to remove variables not significantly associated with outcome. All univariate variables with P values 0.20 were considered for multivariate analysis. A receiver operating characteristic (ROC) curve was generated for prediction of 12-month creatinine 1.5 mg/dl by natural logarithm of duration of renal disease among OLTa patients. Statistical programs used were SAS 8.2 (SAS Institute, Cary, NC) and Stata 8.1 (Stata Corp, College Station, TX). A P value 0.05 is considered significant. Results We identified 56 patients with pretransplant renal dysfunction who underwent OLTa and 13 subjects who received CKLT (Table 1). Among OLTa patients, 15 (27%) had diabetes mellitus, 20 (36%) had hypertension, and 15 (27%) were diagnosed with hepatorenal syndrome. Thirty-two (57%) of OLTa patients did not have HRS. Causes for renal dysfunction (more than 1 identified in 3 subjects) among non-hrs OLTa patients were diuretics (19), acute tubular necrosis (8), glomerulonephritis (2), infection (2), volume depletion (1), hypotension (1), and unknown (2). Mean MELD was Of 34 patients with viral hepatitis, 7 had both hepatitis C and alcohol-related liver disease. Thirteen patients had other causes of liver disease: autoimmune hepatitis (n 3), alpha-1 antitrypsin disease (n 3), hemochromatosis (n 2), Budd-Chiari syndrome (n 2), veno-occlusive disease (n 1), cystic fibrosis (n 1), and polycystic liver disease (n 1). Patients undergoing CKLT as opposed to OLTa had longer duration of pretransplant renal dysfunction (median 18.1 vs. 2.7 weeks, P 0.001). Seventy-five

4 Duration of Renal Dysfunction and OLT 1051 Table 2. Post-OLT Characteristics of OLTa and CKLT Patients Variable OLTa (n 56) n (%) or Median (Interquartile Range) CKLT (n 13) n (%) or Median (Interquartile Range) P Value Cyclosporine use 0 (0%) 0 (0%) Tacrolimus use 51 (91%) 13 (100%) 0.58 Sirolimus use 14 (25%) 2 (15%) month creatinine 1.4 mg/dl ( mg/dL) 1.6 mg/dl ( mg/dL) month creatinine 1.4 mg/dl ( mg/dL) 1.5 mg/dl ( mg/dL) 0.97 RRT ever post-olt 13 (23%) 2 (15%) 0.72 RRT at 6 months 0 (0%) 0 (0%) RRT at 12 months 1 (2%) 0 (0%) percent of OLTa patients had a duration of pretransplant renal dysfunction less than 6.4 weeks, whereas 75% of CKLT patients had elevated creatinine levels for longer than 8.0 weeks. CKLT patients, as opposed to OLTa patients, also had higher serum creatinine (median, 4.0 vs 1.7 mg/dl, P 0.001), and were more likely to have received RRT (62% vs. 7%, P 0.001) before transplantation. RRT in each case was hemodialysis. Duration of RRT for the 4 OLTa patients was 0.14, 0.57, 0.71, and 1.43 weeks. Eight CKLT patients required RRT for a median of 7.6 weeks (range, weeks), P compared to OLTa. Subjects were otherwise similar between the 2 groups. CKLT patients had similar post-olt creatinine compared to OLTa patients, and only one person required RRT by 12-month follow-up (Table 2). Patients were selected for CKLT because transplant nephrologists felt recovery of renal function after OLTa would be unlikely (Table 3). In 3 cases, the primary reason for CKLT was a relatively long duration of acute renal insufficiency. Only 3 patients had end-stage renal disease requiring more than 9 weeks of RRT prior to transplantation (Table 3). We documented biopsy results in 2 subjects (patients 4 and 9). Duration of pretransplant renal disease was a significant predictor of post-olta creatinine (Tables 4 and 5). Multivariate analysis showed that both duration of pretransplant renal disease (P 0.002) and other Table 3. Characteristics of CKLT Patients Patient Reason for CKLT Duration of Renal Disease (weeks) Duration of RRT, If Required (weeks) 1 Long duration of possible HRS End-stage hypertensive nephropathy Multifactorial renal insufficiency and planned nephrectomy 12.4 N/A for renal cell carcinoma at time of transplant 4 Membranoproliferative glomerulopathy 32.3 N/A 5 Long duration of HRS and chronic renal insufficiency Long duration of HRS Multifactorial renal insufficiency and planned nephrectomy for renal cell carcinoma at time of transplant 8 Membranoproliferative glomerulonephritis and diabetic glomerulopathy 9 Membranoproliferative glomerulonephritis 3.4 N/A 10 Polycystic kidney disease Acute renal failure induced by diuretics and uncertain cause 18.1 N/A chronic renal insufficiency 12 NSAID nephropathy 6.1 N/A 13 Lupus nephritis and cyclosporine toxicity Abbreviations: N/A, not applicable; NSAID, nonsteroidal antiinflammatory drug.

5 1052 Campbell et al. Table 4. Predictors of Serum Creatinine 6 Months After OLTa Univariate Analysis Duration of pretransplant renal disease ( )* 0.04 Multivariate Analysis Duration of pretransplant renal disease ( )* Cause of liver disease (compared to other) Viral 0.32 ( ) NASH 0.30 ( ) 0.01 Alcohol 0.39 ( ) Abbreviation: NASH, nonalcoholic steatohepatitis. *Inverse creatinine (mg/dl) per natural logarithm of duration (weeks). Inverse creatinine (mg/dl). cause of liver disease (P 0.01) predicted 6-month post-olta creatinine (Table 4). Pretransplantation creatinine (P 0.001) and bilirubin (P 0.006) were significant predictors of post-olta creatinine at 12 months (Table 5). All other variables analyzed were not associated with post-olta creatinine. We performed ROC analysis for prediction of creatinine 1.5 mg/dl at 12 months after OLTa (Fig. 1). Area under the ROC curve was Using duration of 3.6 weeks as a test threshold, sensitivity was 62.5% with specificity 69.2%, positive predictive value 71.4%, and negative predictive value 60.0%. Our formula for prediction of creatinine (mg/dl) 12 months after OLTa is: 1/(.8595 [1/pretransplant creatinine (mg/dl)] ln[bilirubin (mg/ dl)] ln[duration of renal insufficiency (weeks)] ), (model R , adjusted R ). For example, a patient with 8 weeks duration of renal insufficiency, creatinine 3.0 mg/dl, and bilirubin 5.0 mg/dl would be expected to have a creatinine of 2.0 mg/dl post-olta. We applied this model, derived from OLTa patients, to the 4 CKLT subjects with possible or definite HRS. Predicted 12-month creatinine in the absence of concomitant renal transplanta- Table 5. Predictors of Serum Creatinine 12 Months After OLTa Univariate Analysis Duration of pretransplant renal disease ( )* 0.01 Diabetes mellitus 3.21 ( )* Pretransplant creatinine 0.48 ( ) 0.01 Bilirubin ( ) 0.02 INR 0.22 ( ) 0.02 Multivariate Analysis Duration of pretransplant renal disease ( )* 0.10 Pretransplant creatinine 0.86 ( ) Bilirubin ( ) Abbreviation: INR, international normalized ratio. *Inverse creatinine (mg/dl) per natural logarithm of duration (weeks). Inverse creatinine post-olta (mg/dl) per inverse creatinine pre-olta (mg/dl). Inverse creatinine (mg/dl) per natural logarithm of bilirubin (mg/dl). Inverse creatinine (mg/dl) per natural logarithm of INR.

6 Duration of Renal Dysfunction and OLT 1053 Figure 1. ROC curve for duration of pre-olta renal dysfunction predicting creatinine >1.5 mg/dl 12 months post-olta. Each dot represents a different test threshold. tion was 2.0 mg/dl for all 4 patients (range, mg/dl). Predictors of posttransplantation creatinine were assessed among all patients, forcing the transplantation group (CKLT vs. OLTa) into the multivariate analysis (Table 6). Duration of pretransplant renal disease was a significant predictor of both 6- and 12-month post-transplantation creatinine (P 0.02, P 0.003). After adjusting for duration of pretransplant renal disease and other baseline characteristics, CKLT was associated with significantly lower creatinine at 12 months compared to OLTa. Discussion These data suggest that duration of pretransplant renal dysfunction is an independent predictor of renal function after OLTa. Interestingly, presence or absence of HRS was not associated with post-olta renal outcomes. It has long been recognized that preexisting chronic kidney disease is an important indication for CKLT because pretransplant renal function will not improve post-olta In contrast, HRS has generally portended a better renal prognosis after OLTa. 14,15 However, duration of pretransplant renal dysfunction is an important confounding variable that has not been studied. It is possible that previous studies have demonstrated better renal outcomes for HRS patients because those patients had the shortest durations of renal dysfunction prior to transplantation. Our study suggests that duration, not cause, of pretransplant renal dysfunction is key to predicting creatinine after transplantation. A thorough effort was made in our study to accurately diagnose HRS. Three investigators reviewed all available data and agreed on a renal diagnosis for each patient, following International Ascites Club guidelines. 17 Only 27% of OLTa patients were diagnosed with HRS. It is possible that some cases categorized as other or unknown in fact represented HRS. HRS can be difficult to diagnose and requires exclusion of other causes. Misclassification bias is unlikely to affect our results, because we introduced a third category for cases in which HRS could not be ruled definitively in or out. HRS can be divided into types 1 and 2, principally by the rate at which renal dysfunction progresses. We did not distinguish between types 1 and 2, because our duration of renal dysfunction variable makes such a Table 6. Predictors of Serum Creatinine After Transplantation, Adjusting CKLT vs. OLTa by Baseline Characteristics Creatinine 6 Months Post-OLT CKLT (instead of OLTa) 0.12 ( )* 0.15 Duration of pretransplant renal disease ( ) 0.02 Cause of liver disease (compared to other) Viral 0.20 ( )* 0.02 NASH 0.23 ( )* 0.04 Alcohol 0.28 ( )* 0.01 White race (vs. other) 0.20 ( )* 0.02 Creatinine at 12 Months Post-OLT CKLT (instead of OLTa) 0.21 ( )* 0.01 Duration of pretransplant renal disease ( ) White race (vs. other) 0.28 ( )* RRT pretransplantation 0.19 ( )* 0.03 Abbreviation: NASH, nonalcoholic steatohepatitis. *Inverse creatinine (mg/dl). Inverse creatinine (mg/dl) per natural logarithm of duration (weeks).

7 1054 Campbell et al. distinction redundant. One possible explanation for our results could be that type 1 HRS (shorter duration of renal dysfunction) portends better posttransplantation renal function than type 2 (longer duration of renal dysfunction). Clinically, it is often quite challenging to distinguish HRS from other conditions associated with effective volume contraction. Furthermore, renal dysfunction may be multifactorial. Therefore, duration may be a more user-friendly clinical variable. We chose a creatinine cutoff of 1.5 mg/dl for inclusion in the study. Such a creatinine value is the minimum required to diagnose HRS. 17 Other studies have used the same or lower creatinine cutoffs to define pretransplant renal dysfunction. 5-7 In a patient with endstage liver disease, a creatinine of 1.5 mg/dl signifies more advanced renal impairment than in a patient without liver disease. Cirrhotics have decreased synthesis of creatinine, decreased muscle mass, hyperbilirubinemia, and volume expansion. 16 Formulas such as the Cockcrot-Gault equation and the Modification of Diet in Renal Disease formula use creatinine, age, serum albumin, serum blood urea nitrogen, and/or body weight to estimate glomerular filtration rate. Neither these formulas nor calculation of creatinine clearance from a 24-hour urine collection has been well studied or validated in patients with decompensated cirrhosis. 17 Ideally, renal function can be estimated through the use of inulin, [ 125 I]iothalamate, or 51Cr-EDTA clearance methods, though this is costly and often impractical. After transplantation, only one patient developed end stage renal disease requiring continuing RRT. However, one-half of patients had serum creatinine higher than 1.5 mg/dl, regardless of whether or not they had received a concomitant kidney with a liver. Even moderate renal insufficiency in posttransplantation patients can be significant, because renal dysfunction after transplantation may be progressive, in part because of nephrotoxic immunosuppression regimens, development of hypertension, and diabetes mellitus. Short-term renal dysfunction after transplantation has been associated with long-term poorer prognosis. 1,18 Predictive variables other than duration of pretransplant renal dysfunction were identified. Preoperative serum creatinine and diabetes mellitus are well-known risk factors for the development of renal dysfunction. Multivariable analysis of OLTa patients showed bilirubin and creatinine, 2 components of MELD, were predictive of 12-month creatinine. A prior study identified MELD as a variable associated with postoperative renal dysfunction, similar to our result. 19 Other variables found to be associated with worse post-olt renal function were white race and other cause of liver disease. Both of these findings could potentially be artifactual, given the relatively small numbers of patients who were nonwhite and who had other cause of liver disease. Only 1 patient had polycystic liver disease. Patients undergoing CKLT as opposed to OLTa were different in several key respects. CKLT patients had longer duration of pretransplant renal dysfunction (median, 18.1 vs. 2.7 weeks), higher serum creatinine, and increased requirement for RRT preoperatively. Whereas 75% of OLTa patients had pretransplant renal dysfunction for less than 6.4 weeks, 75% of CKLT patients had elevated creatinine lasting over 8.0 weeks. Postoperatively, the 2 groups of patients were exposed to similar rates of cyclosporine, tacrolimus, and sirolimus usage. Serum creatinine values at 6 and 12 months were also similar between CKLT and OLTa patients. After adjusting for baseline characteristics (including duration of pretransplant renal disease), CKLT patients had better renal function at 12 months (statistical significance not reached at 6 months). These data suggest that in our study population, patients with features predictive of poor renal function after liver transplantation were appropriately selected for CKLT. ROC analysis among OLTa patients showed that duration of renal disease by itself had a moderate ability to predict creatinine 1.5 mg/dl at 12 months posttransplantation (area under ROC curve 0.71). The optimal predictive cutoff was 3.6 weeks. It is noteworthy that all but 1 CKLT patient did have duration of renal disease 3.6 weeks (1 CKLT patient had 3.4 weeks of renal insufficiency). However, we cannot at this time recommend that all patients with duration of renal disease longer than 3.6 weeks undergo CKLT. A creatinine of 1.5 mg/dl 1 year after transplantation is not necessarily high enough to justify concomitant renal transplantation. Instead, long-term studies are needed to determine what creatinine can be tolerated at 1 year post-olta before renal failure becomes likely over the next several years. Because 75% of our OLTa patients had creatinine 1.9 mg/dl or less at 12 months post-olta, we could not perform ROC analyses for higher creatinine outcomes. Furthermore, our area under the ROC curve is not high enough to justify use of duration of renal disease in isolation. Instead, other predictive features, such as height of creatinine and requirement for RRT, may be considered in combination with the duration of renal disease to predict post- OLTa renal function. We applied our multivariable model, derived from OLTa patients, to predict 12-month creatinine posttransplantation in the 4 CKLT subjects with definite or possible HRS. In the absence of renal transplantation,

8 Duration of Renal Dysfunction and OLT 1055 all 4 CKLT patients would have expected creatinine 2.0 mg/dl 12 months post-olta. Although we cannot definitively state how high a 1-year posttransplantation creatinine can be tolerated before CKLT should be performed, a creatinine value in the range of 2-3 mg/dl seems reasonable. In summary, our data suggest that duration of pretransplant renal dysfunction may help identify those patients most in need of CKLT as opposed to OLTa. A threshold duration of renal dysfunction in combination with other clinical variables may be prospectively investigated as an aid to clinical decisionmaking. Acknowledgment The authors thank Mary Kaminski of the Department of Surgery, University of Pennsylvania Health System, for her contributions to this study. References 1. Gonwa TA, Mai ML, Melton LB, Hays SR, Goldstein RM, Levy MF, et al. Renal replacement therapy (RRT) and orthotopic liver transplantation (OLTX): The role of continuous veno-venous hemodialysis. Transplantation 2001;71: Nair S, Verma S, Thuluvath PJ. Pretansplant renal function predicts survival in patients undergoing orthotopic liver transplantation. Hepatology 2002;35: Markmann JF, Markmann JW, Markmann DA, Bacquerizo A, Singer J, Holt CD, et al. Preoperative factors associated with outcome and their impact on resource use in 1148 consecutive primary liver transplants. Transplantation 2001;72: Brown RS, Lombardero M, Lake JR. Outcome of patients with renal insufficiency undergoing liver or liver-kidney transplantation. Transplantation 1996;62: Bilbao I, Charco R, Balsells J, Lazaro JL, Hildalgo E, Llopart L, et al. Risk factors for acute renal failure requiring dialysis after liver transplantation. Clin Transplant 1998; 12: Lafayette RA, Pare G, Schmid CH, King AJ, Rohrer RJ, Nasraway SA. Pretransplant renal dysfunction predicts poorer outcome in liver transplantation. Clin Nephrol 1997;48: Pawarode A, Fine DM, Thuluvath PJ. Independent risk factors and natural history of renal dysfunction in liver transplant recipients. Liver Transpl 2003;9: Margreiter R, Konigsrainer A, Spechtenhauser B, Ladurner R, Pomarolli A, Hormann Ch, et al. Our experience with combined liver-kidney transplantation: An update. Transplant Proc 2002; 34: Lang M, Neumann U, Kahl A, Steinmuller T, Settmacher U, Neuhas P. Long-term outcome of 27 patients after combined liver-kidney transplantation. Transplant Proc 2001;33: Ammor M, Creput C, Durrbach A, Samuel D, Von Ey, F, Hiesse C, et al. Mortality and long term outcome of combined liver and kidney transplantations. Transplant Proc 2001;33: Gatti S, Arru M, Reggiani P, Rossi G, Tarantino A, Berardinelli L, et al. Combined liver and kidney transplantation: A singlecenter experience. Transplant Proc 2002;34: Torregrosa JV, Inigo P, Navasa M, Rimola A, Grande L, Oppenheimer F. Combined liver-kidney transplantation: Our experience. Transplant Proc 1999;31: Chui AKK, DeLeon C, Rao ARN, Verran DJ, Pathania OP, McCaughan GW, et al. Single-center experience of combined liver and kidney transplantation. Transplant Proc 1998;30: Seu P, Wilkinson AH, Shaked A, Busuttil RW. The hepatorenal syndrome in liver transplant recipients. Am Surg 1991;57: Gonwa TA, Morris CA, Goldstein RM, Husberg BS, Klintmalm GM. Long-term survival and renal function following liver transplantation in patients with and without hepatorenal syndrome experience in 300 patients. Transplantation 1991;51: Davis CL, Gonwa TA, Wilkinson AH. Identification of patients best suited for combined liver-kidney transplantation: Part II. Liver Transpl 2002;8: Arroyo V, Gines P, Gerbes AL, Dudley FJ, Gentilini P, Laffi G, et al. Definition and diagnostic criteria of refractory ascites and hepatorenal syndrome in cirrhosis. Hepatology 1996;23: Velidedeoglu E, Bloom RD, Crawford MD, Desai NM, Campos L, Abt PL, et al. Early kidney dysfunction post liver transplantation predicts late chronic kidney disease. Transplantation 2004; 77: Sanchez EQ, Gonwa TA, Levy MR, Goldstein RM, Mai ML, Hays SR, et al. Preoperative and perioperative predictors of the need for renal replacement therapy after orthotopic liver transplantation. Transplantation 2004;78: