Begin typing your search above and press return to search.
Volume: 19 Issue: 3 March 2021


Pretransplant Use of the Chronic Kidney Disease Epidemiology Collaboration Equation (CKD-EPI) to Estimate Glomerular Filtration Rate Predicts Outcomes in Liver Transplant Recipients


Objectives: Kidney dysfunction is common in liver transplant candidates and is a well-established predictor of increased mortality after liver transplant. However, the best method for determination of the glomerular filtration rate before liver transplant remains unclear.
Materials and Methods: We analyzed the performance of the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation and the Modification of Diet in Renal Disease (MDRD) Study equation, before liver transplant, compared with radionuclide glomerular filtration rate and examined the association of the 2 equations with a composite outcome of stage 4 chronic kidney disease, initiation of chronic dialysis, or patient death.
Results: We studied 426 consecutive adult liver transplant recipients from 1990 to 2014. The correlation coefficient of the radionuclide glomerular filtration rate with the Chronic Kidney Disease Epidemiology Collaboration equation was 0.61 and with the Modification of Diet in Renal Disease Study equation was 0.58. The Modification of Diet in Renal Disease Study equation showed a bias of -4.7 mL/min and precision of 32.9 mL/min, whereas the Chronic
Kidney Disease Epidemiology Collaboration equation showed a bias of -11.1 mL/min but was more precise (28.1 mL/min). Only the Chronic Kidney Disease Epidemiology Collaboration equation remained significantly associated with the composite outcome in the multivariable analysis.
Conclusions: The use of the Chronic Kidney Disease Epidemiology Collaboration equation in the period before liver transplant provided independent prognostic information regarding long-term outcomes after liver transplant.

Key words : CKD-EPI equation, Liver transplantation, MDRD equation, Radionuclide glomerular filtration rate


Preexisting kidney dysfunction is a well-established predictor of increased morbidity and mortality after liver transplant (LTx).1,2 There are several mechanisms involved in kidney disease that may lead to kidney failure in patients with liver cirrhosis.3,4 In addition, patients may present with other comorbid conditions such as diabetes and/or hypertension,5,6 which are well-known risk factors for chronic kidney disease (CKD)7,8 and may affect LTx outcomes.5,6

A systematic review showed that pre-LTx impairment in kidney function is a risk factor for CKD post-LTx.9 Both duration and severity of kidney dysfunction prior to LTx have been associated with postoperative progression of kidney disease.10,11 Therefore, the evaluation of renal function pre-LTx is of utmost importance.

Because the best method for determination of the glomerular filtration rate (GFR) pre-LTx remains unclear, we examined both the correlation and perfor­mance of the Modification of Diet in Renal Disease Study Group equation (4 variables, MDRD-4)12 and the Chronic Kidney Disease Epidemiology Collabo­ration equation (CKD-EPI)13 with the radionuclide GFR (rGFR). We also assessed the effect of the pre-LTx rGFR and the estimated GFR (eGFR) equations as predictors of post-LTx progression to CKD and mortality.

Materials and Methods

Study design
This was a single-center, retrospective cohort study that included consecutive LTx candidates between January 1990 and December 2014. Data were extracted from a prospectively derived transplant database at the McGill University Health Center. Patient race/ethnicity was categorized as African Canadian, White, and Other based on self-reporting by patients. The study protocol was approved by the Institutional Research Ethics Board. We studied 426 consecutive LTx recipients, after exclusion of patients with combined organ transplants (n = 43), retransplants (n = 213), and without measured GFR or serum creatinine within 1 year of transplant (n = 190), as well as patients who were on dialysis pre-LTx at the time of evaluation for LTx (n = 6). The GFR for all patients was estimated with 2 serum creatinine-based equations (MDRD-4 and CKD-EPI) and was measured with rGFR. The rGFR was measured by technetium-99m-labeled diethylenetriaminepentaaccetic acid (99mTc-DTPA) plasma clearance (at 1 and 3 hours) within 1 year pre-LTx. We classified CKD stages according to the Kidney Disease Improving Global Outcomes guidelines (KDIGO 2012). Baseline eGFR was calculated with MDRD-4 and CKD-EPI within 3 months pre-LTx, using the lowest serum creatinine value. Serum creatinine was measured with the modified Jaffe reaction. The assay was calibrated to the isotope dilution mass spectrometry standard in 2008. As previously reported, the effect of this method in clinical practice is minimal.14

Statistical analyses
Results for continuous variables are expressed as means ± SD. The correlation of MDRD and CKD-EPI equations with rGFR was evaluated using the Pearson correlation test. The t test, chi-square test, Mann-Whitney test, logistic regression analysis, and Cox regression were used where appropriate. Paired t test was used to compare differences in the mean values. The performance of both eGFR equations was compared with rGFR using the Bland-Altman plot to identify bias, precision, and accuracy.15 Bias expresses the systematic deviation from the mean rGFR and was given by the mean difference between equation-based GFR estimates and rGFR measurements. Precision of the estimates is determined as the SD of the mean difference between eGFR and rGFR. Accuracy incor­porates both bias and precision and was calculated as the percentage of the GFR estimates within 30% deviation of the measured GFR. The combined root mean square error was used to capture both bias and precision and is calculated as the square root of the following sum: (mean difference between estimated and measured GFR)2 + (SD of the difference)2. We examined the association of each GFR measurement with a composite clinical outcome, including CKD stage 4 (GFR <30 mL/min/1.73 m2), need for chronic dialysis, or death in univariable and multivariable analysis using adjusted Cox proportional hazards models. Statistical analyses were performed using SAS version 9.1 (SAS Institute). Two-sided P value of .05 was deemed statistically significant.


There were 426 LTx recipients (all from deceased donors) eligible for inclusion into our study from 1990 to 2014. Baseline characteristics are shown in Table 1. The median follow-up was 5.7 years. The mean age at the time of listing was 56.4 ± 10.5 years; 286 were men (67%), 340 were White (80%), and 16 (4%) were African Canadians. The mean body mass index (BMI, (calculated as weight in kilograms divided by height in meters squared) was 27.9 ± 5.8. The mean Model for End-Stage Liver Disease score, with sodium, was 22.2 ± 8.5. Hepatitis C was the more prevalent primary disease (29%), followed by alcoholic liver disease (19%), nonalcoholic steatohepatitis (11%), and hepatocellular carcinoma (11%). The pre-LTx median serum creatinine was 78 μmol/L (range, 64-102 μmol/L), median rGFR was 92 mL/min/1.73 m2 (interquartile range [IQR], 69-116 mL/min/1.73 m2), CKD-EPI eGFR was 89 mL/min/1.73 m2 (IQR, 64-102 mL/min/1.73 m2), and MDRD-4 eGFR was 87 mL/min/1.73 m2 (IQR, 65-110 mL/min/1.73 m2) (Table 1). The correlation coefficient of the rGFR with the CKD-EPI and the MDRD-4 equations was 0.61 (P < .001) and 0.58 (P < .001), respectively (Figure 1). On average, the CKD-EPI equation underestimated the rGFR by 11.1 mL/min, whereas the MDRD equation underestimated the rGFR by 4.7 mL/min. The combined root mean square error showed that the CKD-EPI equation can perform better than the MDRD-4 equation (Figure 2).

Univariable and multivariable Cox regression models were used to identify variables associated with patient clinical outcomes, including CKD stage 4, end-stage renal disease requiring chronic dialysis, and patient survival. Of 183 patients that reached the composite outcome, 42 had CKD stage 4 or required chronic dialysis, and 162 died. The values for rGFR and eGFR were significantly associated with the composite outcome in univariable analysis (Table 2). However, after adjusting for potential confounders (age, race, BMI, etiology of end-stage liver disease, and preexisting diabetes) in a backward selection approach, only the CKD-EPI equation was significantly associated with the composite outcome, with 9% decrease in the relative hazard for the composite outcome for every 10-mL/min increase of eGFR (Table 3).

The causes of death are shown in Table 4.


We found that the pre-LTx CKD-EPI and MDRD-4 equations had a modest correlation with rGFR and that the CKD-EPI equation independently predicted the composite outcome of CKD stage 4, initiation of chronic dialysis, or patient death post-LTx, compared with the rGFR and the MDRD-4 equation. Other studies have shown that the assessment of kidney function in the pre-LTx evaluation, as well as in the posttransplant follow-up, predicts the risk of patient survival.9,16-18 Given that there is no gold standard of GFR measurement in LTx candidates and recipients, we used the rGFR as the reference method, based on its good correlation with the inulin clearance.19 The modest correlation that we observed between the CKD-EPI and the MDRD-4 equations with the rGFR may be explained by the effect of ascites or volume excess on serum creatinine measurements in patients with liver diseases. This was illustrated by Xie and colleagues, who reported that CKD-EPI was more accurate than the 99mTc-DTPA in reference to measured GFR.20

Although the CKD-EPI and MDRD equations underestimated the incidence of CKD stage 4 or end-stage renal disease that required chronic dialysis in our study cohort, similar to previous research,21,22 the CKD-EPI equation had greater precision and reliability compared with the MDRD equation, especially for patients with rGFR >60 mL/min/1.73 m2. In fact, the poor performance of GFR-estimating equations and the overestimation of the GFR by measured creatinine clearance in patients with liver disease is well known.23 This could be caused by changes in the muscle mass found in patients with liver disease and cirrhosis.24 Therefore, equations developed in kidney and heart transplant cohorts may not be applicable to LTx candidates.

It is known that there are differences in precision and bias in the MDRD24,25 and CKD-EPI equations,24 based on measured GFR, which is also the case in our study. Cystatin C is less affected by muscle volume26; therefore, cystatin C may be an ideal biomarker with which to estimate GFR in patients with cirrhosis or to predict outcomes in cirrhotic patients with ascites. De Souza and colleagues evaluated the performance of several CKD equations (CKD-EPI for serum creatinine, CKD-EPI for cystatin C, and CKD-EPI for serum creatinine and cystatin C), the MDRD equation (4 and 6 variables), and the Hoek formula with respect to inulin clearance in cirrhotic patients pre-LTx.25 They reported that serum creatinine-based equations were not reliable, whereas cystatin C-based equations showed a better performance, including the CKD-EPI cystatin C equation, which was the best equation irrespective of the ascites severity or renal dysfunction. In a recent publication, Mauro and colleagues reported that an eGFR <60 mL/min/1.73 m2 from the CKD-EPI equation for serum creatinine and cystatin C at the time of wait list placement was a predictor of the need for renal replacement therapy during the first month post-LTx.27

This study has some limitations, including the fact that eGFR and rGFR were not always measured on the same day, and we used neither inulin clearance (which is considered the gold standard method for GFR determination) nor cystatin C to estimate GFR. Finally, most of the LTx recipients were White with preserved kidney function, and so it was difficult to generalize the results to other LTx candidates.

The correlation of the CKD-EPI and MDRD-4 equations with rGFR is suboptimal. However, in our patient population, the use of the CKD-EPI equation in the pre-LTx setting predicted the progression of CKD and patient death post-LTx. On the basis of our results, it is critical to assess the potential factors that may affect eGFR and how the eGFR may perform as a predictor of kidney failure and patient survival in different subpopulations in the pre-LTx and post-LTx periods. Determination of these factors may contribute to recalibration of existing equations for estimation of GFR in LTx candidates.


  1. Nair S, Verma S, Thuluvath PJ. Pretransplant renal function predicts survival in patients undergoing orthotopic liver transplantation. Hepatology. 2002;35(5):1179-1185. doi:10.1053/jhep.2002.33160
    CrossRef - PubMed
  2. Molnar MZ, Joglekar K, Jiang Y, et al. Association of pretransplant renal function with liver graft and patient survival after liver transplantation in patients with nonalcoholic steatohepatitis. Liver Transpl. 2019;25(3):399-410. doi:10.1002/lt.25367
    CrossRef - PubMed
  3. Gines P, Schrier RW. Renal failure in cirrhosis. N Engl J Med. 2009;361(13):1279-1290. doi:10.1056/NEJMra0809139
    CrossRef - PubMed
  4. Weber ML, Ibrahim HN, Lake JR. Renal dysfunction in liver transplant recipients: evaluation of the critical issues. Liver Transpl. 2012;18(11):1290-1301. doi:10.1002/lt.23522
    CrossRef - PubMed
  5. Velidedeoglu E, Bloom RD, Crawford MD, et al. Early kidney dysfunction post liver transplantation predicts late chronic kidney disease. Transplantation. 2004;77(4):553-556. doi:10.1097/
    CrossRef - PubMed
  6. Glowczynska R, Galas M, Witkowska A, et al. The pre-transplant profile of cardiovascular risk factors and its impact on long-term mortality after liver transplantation. Ann Transplant. 2018;23:591-597. doi:10.12659/AOT.908771
    CrossRef - PubMed
  7. Kowalski A, Krikorian A, Lerma EV. Diabetes and chronic kidney disease. Dis Mon. 2015;61(9):378-386. doi:10.1016/j.disamonth.2015.07.002
    CrossRef - PubMed
  8. Gargiulo R, Suhail F, Lerma EV. Hypertension and chronic kidney disease. Dis Mon. 2015;61(9):387-395. doi:10.1016/j.disamonth.2015.07.003
    CrossRef - PubMed
  9. Fabrizi F, Dixit V, Martin P, Messa P. Pre-transplant kidney function predicts chronic kidney disease after liver transplant: meta-analysis of observational studies. Dig Dis Sci. 2011;56(5):1282-1289. doi:10.1007/s10620-010-1529-2
    CrossRef - PubMed
  10. Pham PT, Pham PC, Wilkinson AH. Renal function outcomes following liver transplantation and combined liver-kidney transplantation. Nat Clin Pract Nephrol. 2007;3(9):507-514. doi:10.1038/ncpneph0574
    CrossRef - PubMed
  11. Campbell MS, Kotlyar DS, Brensinger CM, et al. Renal function after orthotopic liver transplantation is predicted by duration of pretransplantation creatinine elevation. Liver Transpl. 2005;11(9):1048-1055. doi:10.1002/lt.20445
    CrossRef - PubMed
  12. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med. 1999;130(6):461-470. doi:10.7326/0003-4819-130-6-199903160-00002
    CrossRef - PubMed
  13. Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604-612. doi:10.7326/0003-4819-150-9-200905050-00006
    CrossRef - PubMed
  14. Killeen AA, Ashwood ER, Ventura CB, Styer P. Recent trends in performance and current state of creatinine assays. Arch Pathol Lab Med. 2013;137(4):496-502. doi:10.5858/arpa.2012-0134-CP
    CrossRef - PubMed
  15. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res. 1999;8(2):135-160. doi:10.1177/096228029900800204
    CrossRef - PubMed
  16. Uguen T, Jezequel C, Ropert M, et al. Pretransplant renal function according to CKD-EPI cystatin C equation is a prognostic factor of death after liver transplantation. Liver Int. 2016;36(4):547-554. doi:10.1111/liv.12989
    CrossRef - PubMed
  17. Sharma P, Schaubel DE, Guidinger MK, Merion RM. Effect of pretransplant serum creatinine on the survival benefit of liver transplantation. Liver Transpl. 2009;15(12):1808-1813. doi:10.1002/lt.21951
    CrossRef - PubMed
  18. Cantarovich M, Tchervenkov J, Paraskevas S, et al. Early changes in kidney function predict long-term chronic kidney disease and mortality in patients after liver transplantation. Transplantation. 2011;92(12):1358-1363. doi:10.1097/TP.0b013e3182384aff
    CrossRef - PubMed
  19. Gunasekera RD, Allison DJ, Peters AM. Glomerular filtration rate in relation to extracellular fluid volume: similarity between 99mTc-DTPA and inulin. Eur J Nucl Med. 1996;23(1):49-54. doi:10.1007/BF01736989
    CrossRef - PubMed
  20. Xie P, Huang JM, Liu XM, Wu WJ, Pan LP, Lin HY. (99m)Tc-DTPA renal dynamic imaging method may be unsuitable to be used as the reference method in investigating the validity of CDK-EPI equation for determining glomerular filtration rate. PLoS One. 2013;8(5):e62328. doi:10.1371/journal.pone.0062328
    CrossRef - PubMed
  21. Murata K, Baumann NA, Saenger AK, Larson TS, Rule AD, Lieske JC. Relative performance of the MDRD and CKD-EPI equations for estimating glomerular filtration rate among patients with varied clinical presentations. Clin J Am Soc Nephrol. 2011;6(8):1963-1972. doi:10.2215/CJN.02300311
    CrossRef - PubMed
  22. Stevens LA, Schmid CH, Greene T, et al. Comparative performance of the CKD Epidemiology Collaboration (CKD-EPI) and the Modification of Diet in Renal Disease (MDRD) Study equations for estimating GFR levels above 60 mL/min/1.73 m2. Am J Kidney Dis. 2010;56(3):486-495. doi:10.1053/j.ajkd.2010.03.026
    CrossRef - PubMed
  23. Francoz C, Prie D, Abdelrazek W, et al. Inaccuracies of creatinine and creatinine-based equations in candidates for liver transplantation with low creatinine: impact on the model for end-stage liver disease score. Liver Transpl. 2010;16(10):1169-1177. doi:10.1002/lt.22128
    CrossRef - PubMed
  24. Gonwa TA, Jennings L, Mai ML, Stark PC, Levey AS, Klintmalm GB. Estimation of glomerular filtration rates before and after orthotopic liver transplantation: evaluation of current equations. Liver Transpl. 2004;10(2):301-309. doi:10.1002/lt.20017
    CrossRef - PubMed
  25. De Souza V, Hadj-Aissa A, Dolomanova O, et al. Creatinine- versus cystatine C-based equations in assessing the renal function of candidates for liver transplantation with cirrhosis. Hepatology. 2014;59(4):1522-1531. doi:10.1002/hep.26886
    CrossRef - PubMed
  26. Shlipak MG, Mattes MD, Peralta CA. Update on cystatin C: incorporation into clinical practice. Am J Kidney Dis. 2013;62(3):595-603. doi:10.1053/j.ajkd.2013.03.027
    CrossRef - PubMed
  27. Mauro E, Crespo G, Martinez-Garmendia A, et al. Cystatin C and sarcopenia predict acute on chronic liver failure development and mortality in patients on the liver transplant waiting list. Transplantation. 2020;104(7):e188-e198. doi:10.1097/TP.0000000000003222
    CrossRef - PubMed

Volume : 19
Issue : 3
Pages : 231 - 236
DOI : 10.6002/ect.2020.0557

PDF VIEW [248] KB.

From the 1Division of Nephrology and Multi-Organ Transplant Program, McGill University Health Center, Montreal, Quebec, Canada; and the 2Division of Gastroenterology, McGill University Health Center, Montreal, Quebec, Canada
Acknowledgements: We thank Dr. Andrea Herrera-Gayol for critical review and editing of the manuscript and Dr. David Blank for comments about the measurement of serum creatinine.
This research was supported by personal funding from M. Cantarovich. Other than described above, the authors have not received any funding or grants in support of the presented research or for the preparation of this work and have no further declarations of potential conflicts of interest.
Author contributions: A. Alqallaf reviewed the literature and cowrote the manuscript. A. Alam, RS-P, PG, and MC codesigned the study and cowrote the manuscript.
Current affiliation of Ahmed Alqallaf: Division of Nephrology, Jaber Al-Ahmed Hospital, Kuwait. Current affiliation of Peter Ghali: Division of Gastroenterology, University of Florida, Jacksonville, FL, USA.
Corresponding author: Marcelo Cantarovich, Division of Nephrology and Multi-Organ Transplant Program, McGill University Health Center, Montreal, Quebec, Canada