- Email: [email protected]
Histologic Assessment of Donor Kidneys and Graft Outcome: Multivariate Analyses
Histologic Assessment of Donor Kidneys and Graft Outcome: Multivariate Analyses L.F. Arias, J. Blanco, A. Sanchez-Fructuoso, D. Prats, E. Duque, M. Sáiz-Pardo, J. Ruiz, and A. Barrientos ABSTRACT We performed a multivariate analysis to evaluate the importance of histologic parameters in donor kidney biopsies as predictors of graft outcome. Methods. Wedge protocol biopsies from a single center were analyzed for glomerulosclerosis (GS), interstitial fibrosis (IF), tubular atrophy (TA), arteriosclerosis (AS), and arteriolar hyalinosis (AH). Alterations were quantified as percentage (GS, IF) or semiquantified according to Banff criteria (IF, TA, AS, AH). We calculated creatinine clearance (CrCl) at 1, 2, and 3 years posttransplant. Donor data included age, gender, and type: non– heart-beating donor or brain dead donors. Recipient data included age, gender, cold ischemia time, number of HLA mismatches, peak level of the panel reactive antibody (PRA), number of acute rejection episodes (ARE), and presence or absence of cytomegalovirus (CMV) disease. Univariate and multivariate analyses were performed. Follow-up range was 1 to 4.2 years. Results. GS, IF, TA, and AH were associated with graft survival in the multivariate analysis. The histologic parameters were associated with CrCl at several posttransplant time intervals, but the significance of association was lost in the multivariate analysis. Donor age showed a better correlation with graft function. In the univariate analyses adjusting for donor age, only IF and AH were associated with graft function. Conclusions. Histologic parameters showed a modest association with graft function. In our study, donor age is the better predictor of graft function. IF and AH may be similar to or better than GS as predictors of graft outcome.
M
ORPHOLOGIC factors seemingly impact graft outcomes. Some studies have claimed that glomerulosclerosis (GS) is a good histologic parameter to assess donor kidney biopsies,1,2 while others have claimed that interstitial scarring or vascular changes show a better correlation with graft function.3,4 Other works have reported no correlation between morphologic parameters and graft outcomes.5,6 We performed a histologic evaluation and a multivariate analysis searching for an association between histologic parameters and graft outcome, controlling for confounding factors. MATERIALS AND METHODS Case selection was based on the availability of a donor biopsy and posttransplant follow-up for ⱖ1 year (follow-up range 1 to 4.2 years). Double renal transplants were excluded. Donor data included age, gender, and donor type; non– heart-beating donor
[NHBD] or brain dead. Recipient data included age, gender, cold ischemia time, number of HLA mismatches, peak level of the panel reactive antibody (PRA), number of acute rejection episodes (ARE), and presence of cytomegalovirus (CMV) disease. Graft function was evaluated by creatinine clearance (CrCl) calculated using the Cockcroft-Gault formula at 1, 2, and 3 years after transplantation. Graft failure was defined as return to chronic dialysis. Biopsy specimens were wedge biopsies obtained before the transplantation. Seven to 14 slides were stained with hematoxylineosin, periodic acid-Schiff, and trichrome. GS was registered as a percentage. Interstitial fibrosis (IF), tubular atrophy (TA), arterioFrom the Department of Pathology, University of Antioguia, Medellin Antiquia, Colombia, and Hospital Clínico San Carlos, Madrid, Spain. Address reprint requests to Dr Luis Arias, Department of Pathology, University of Antioquia, Carrera SID No. 62-29, Medellin, Antioquia, Colombia. E-mail: [email protected]
0041-1345/07/$–see front matter doi:10.1016/j.transproceed.2007.01.085
© 2007 by Elsevier Inc. All rights reserved. 360 Park Avenue South, New York, NY 10010-1710
1368
Transplantation Proceedings, 39, 1368 –1370 (2007)
HISTOLOGIC OUTCOMES ASSESSMENT sclerosis (AS), and arteriolar hyalinosis (AH) were graded from 0 to 3 using definitions according to the Banff scheme on allograft pathology. IF was also quantified as a percentage. Immunosuppressive therapy included tacrolimus or cyclosporine plus steroids in all recipients. Other variables were recipient age, delayed graft function, and donor death cause. They did not show any association with graft outcome in the statistical models, so they were excluded in the analyses.
Statistical Analysis All data are expressed as mean values ⫾ SD and, where indicated, as ranges. Unpaired t test or 2 test was used to compare groups. The correlation between variables was assessed using the method of Pearson. The relationship of donor and recipient age, GS, IF, and other continuous variables to CrCl at several time points after transplantation was assessed using simple and multiple linear regression. Multiple logistic regression was used for association of donor and recipient data and graft function. Graft survival was censored for patient death with function. Univariate analyses for graft survival were performed using the Kaplan-Meier method. Cox proportional hazards model was used for the association between donor and recipient data, and graft survival (multivariate analyses). In recipients with graft loss, CrCl was set to the minimal value of 10 mL/min. P values ⬍.05 were considered significant in two-tailed tests. All analysis was done using SPSS software, version 11.5 (SPSS Inc, Chicago, Ill).
RESULTS
The study included 214 kidney biopsies (214 recipients) from 129 donors. Mean donor age was 42.2 ⫾ 16.6 years (range, 2 to 79) with 74.2% men. Overall mean recipient age was 47.7 ⫾ 13.5 years (range, 19 to 71) with 64% men. The main causes of donor death were cardiac disease (36.2%), cerebrovascular disease (31.9%), and trauma (25.7%). There were 58 (45%) NHBD and 71 (55%) brain dead donors. The mean number of glomeruli was 80.9 ⫾ 42.0 (range, 7 to 222) with 96.9% of biopsies containing ⬎20 glomeruli. GS was 6.8 ⫾ 7.8% (range, 0 to 33). In 18 (8.4%) recipients the graft had ⬎20% GS. IF was evident in 158 (73.8%) biopsies. Any grade of TA, AS or AH was identified in 139 (65%), 123 (57.5%), or 129 (60.3%) cases, respectively. A high correlation was observed among the histologic parameters, demonstrating them to be mutually interdependent (r range, 0.51 to 0.97; P ⬍ .001). All histologic variables were significantly associated with donor age (r range, 0.50 to 0.57; P ⬍ .001). In donors whose death was due to cerebrovascular or cardiac diseases, there were more severe histologic alterations. Histologic Parameters and Graft Outcome
Fifteen recipients lost their grafts in the first 3 posttransplant years. GS ⬎20% (log-rank, 7.9; P ⫽ .005), IF moderate and severe (Banff grade 2 to 3; log-rank, 10.4; P ⫽ .001), TA moderate or severe (Banff grade 2 to 3; log-rank, 8.1; P ⫽ .004), and AH moderate or severe (Banff grade 2 to 3; log-rank 14.2; P ⬍ .001) were significantly associated with graft loss. The significant association persisted after adjust-
1369 Table 1. Graft survival: Multivariate Analyses (Cox) Variables
P
Exp(B)
95% CI
Glomerulosclerosis (⬎20%) Interstitial fibrosis (grade 2–3) Tubular atrophy (grade 2–3) Arteriosclerosis (grade 2–3) Arteriolar hyalinosis (grade 2–3) Donor age* Acute rejection episodes (n) HLA mismatches (0–6) CMV disease Cold ischemia time Panel reactive antibody
.008 .003 .011 .051 .000 .384 .183 .409 .293 .129 .162
4.49 4.98 4.12 2.11 2.77 1.02 1.67 .83 .52 .90 1.01
1.4–13.7 1.7–14.3 1.4–12.2 .9–5.9 1.6–4.8 .9–1.7 .8–3.6 .5–1.3 .2–1.8 .8–1.0 1.0–1.03
*Although donor age did not show statistical independent association with graft loss, it was the variable more associated with graft function in the first 3 posttransplant years.
ing for donor age. AS was not independently associated with graft outcome. In the multivariate analysis for graft outcome, GS, IF, TA, and AH were independently associated with graft loss, controlling for donor age, AREs, HLA mismatches, CMV disease, cold ischemia time, and PRA (Table 1). Histologic parameters correlated with CrCl at 1, 2, and 3 posttransplant years, but, controlling for donor age only, IF and AH continue to show significant associations. In the multivariate analysis for graft function (CrCl at 1, 2, and 3 posttransplant years), no independent association was found between histologic variables and function, but increased donor age, number of acute rejection episodes, and CMV disease were independently associated with worse graft function. Graft outcome was not statistically different among NHBD and brain dead donors. DISCUSSION
Previous works have suggested that worse graft survival and function is associated with kidneys showing redused functional renal mass, chronic tubulointerstitial alterations, and arterial and/or arteriolar chronic lesions.1– 4,7 Kidney donor biopsy is more frequently performed for organs from older donors. Donor age is a known risk factor for allograft outcome, but it is not clear whether the histologic study provides additional prognostic information. Some studies have reported a poor correlation between histologic parameters and graft outcome,5,6 while others report an association with donor age.1– 4,8 In the present work we searched for an independent association between histologic parameters and graft survival/function. The initially formulated question was: is age the main donor risk factor, and do morphologic parameters provide additional predictor factors? Because it is clear that severely damaged kidneys are not transplanted, the question is the impact of mild or moderate chronic alterations. Our work has shown a modest association between histologic lesions and graft outcomes; donor age was a better predictor. The association between morphologic parameters and graft survival was clearer than the associa-
1370
tion with graft function. The explanation for this result is not evident. It is possible that more extensive chronic changes impact on outcome earlier. Upon multivariate analysis for survival, there was a low number of events: 15. Thus, there was a low reliability of results in a multivariate analysis. IF and AH were the microscopic variables with more significant P values for association with graft survival. In other works, morphologic parameters variably correlate with posttransplant graft outcome; the differences in results can be due to the methodology of the studies, the number of patients, the categorization of variables, a limitation to biopsies with only mild histologic changes, and lack of control for confounding factors. IF, GS, or AH have variably been identified as critical parameters in several studies. In the present work we found that IF and AH were similar to or better than GS as predictors of graft outcome. Thus, it is important to evaluate vascular and interstitial parameters as precisely as GS in a donor renal biopsy. Although our work showed donor age to be a better donor predictor, this statement did not mean that the donor biopsy is not important. Several cases were discarded because severe histologic changes, and the microscopic studies were decisive. Furthermore knowledge of the baseline morphologic status of an organ is important for subsequent examination and diagnosis in graft biopsies. In conclusion, in the present study, donor age was a better variable to predict graft outcome. Histologic param-
ARIAS, BLANCO, SANCHEZ-FRUCTUOSO ET AL
eters showed a modest association with graft function, but microscopic evaluation provided important information in the decision whether to transplant a kidney.
REFERENCES 1. Escofet X, Osman H, Griffiths DFR, et al: The presence of glomerular sclerosis at time zero has a significant impact on function after cadaveric renal transplantation. Transplantation 75:344, 2003 2. Randhawa PS, Minervini MI, Lombardero M, et al: Biopsy or marginal donor kidneys: Correlation of histologic findings with graft dysfunction. Transplantation 69:1352, 2000 3. Leunissen KM, Bosman FT, Nieman FH, et al: Amplification of the nephrotoxic effect of cyclosporine by preexistent chronic histological lesions in the kidney. Transplantation 48:590, 1989 4. Seron D, Carrera M, Grino JM, et al: Relationship between donor renal interstitial surface and post-transplant function. Nephrol Dial Transplant 8:539, 1993 5. Ratner LE, Joseph V, Zibari G, et al: Transplantation from hypertensive cadaveric donbors. Transplant Proc 27:989, 1995 6. Nyberg G, Hedman L, Blohme I, et al: Morphologic findings in baseline kidney biopsies from living related donors. Transplant Proc 24:355, 1979 7. Karpinski J, Lajoie G, Cattran D, et al: Outcome of kidney transplantation from high-risk donors is determined by both structure and function. Transplantation 67:1162, 1999 8. Pokorná E, Vitko S, Chadimova M, et al: Proportion of glomerulosclerosis in procurement wedge renal biopsy cannot alone discriminate for acceptance of marginal donors. Transplantation 69:36, 2000
M
ORPHOLOGIC factors seemingly impact graft outcomes. Some studies have claimed that glomerulosclerosis (GS) is a good histologic parameter to assess donor kidney biopsies,1,2 while others have claimed that interstitial scarring or vascular changes show a better correlation with graft function.3,4 Other works have reported no correlation between morphologic parameters and graft outcomes.5,6 We performed a histologic evaluation and a multivariate analysis searching for an association between histologic parameters and graft outcome, controlling for confounding factors. MATERIALS AND METHODS Case selection was based on the availability of a donor biopsy and posttransplant follow-up for ⱖ1 year (follow-up range 1 to 4.2 years). Double renal transplants were excluded. Donor data included age, gender, and donor type; non– heart-beating donor
[NHBD] or brain dead. Recipient data included age, gender, cold ischemia time, number of HLA mismatches, peak level of the panel reactive antibody (PRA), number of acute rejection episodes (ARE), and presence of cytomegalovirus (CMV) disease. Graft function was evaluated by creatinine clearance (CrCl) calculated using the Cockcroft-Gault formula at 1, 2, and 3 years after transplantation. Graft failure was defined as return to chronic dialysis. Biopsy specimens were wedge biopsies obtained before the transplantation. Seven to 14 slides were stained with hematoxylineosin, periodic acid-Schiff, and trichrome. GS was registered as a percentage. Interstitial fibrosis (IF), tubular atrophy (TA), arterioFrom the Department of Pathology, University of Antioguia, Medellin Antiquia, Colombia, and Hospital Clínico San Carlos, Madrid, Spain. Address reprint requests to Dr Luis Arias, Department of Pathology, University of Antioquia, Carrera SID No. 62-29, Medellin, Antioquia, Colombia. E-mail: [email protected]
0041-1345/07/$–see front matter doi:10.1016/j.transproceed.2007.01.085
© 2007 by Elsevier Inc. All rights reserved. 360 Park Avenue South, New York, NY 10010-1710
1368
Transplantation Proceedings, 39, 1368 –1370 (2007)
HISTOLOGIC OUTCOMES ASSESSMENT sclerosis (AS), and arteriolar hyalinosis (AH) were graded from 0 to 3 using definitions according to the Banff scheme on allograft pathology. IF was also quantified as a percentage. Immunosuppressive therapy included tacrolimus or cyclosporine plus steroids in all recipients. Other variables were recipient age, delayed graft function, and donor death cause. They did not show any association with graft outcome in the statistical models, so they were excluded in the analyses.
Statistical Analysis All data are expressed as mean values ⫾ SD and, where indicated, as ranges. Unpaired t test or 2 test was used to compare groups. The correlation between variables was assessed using the method of Pearson. The relationship of donor and recipient age, GS, IF, and other continuous variables to CrCl at several time points after transplantation was assessed using simple and multiple linear regression. Multiple logistic regression was used for association of donor and recipient data and graft function. Graft survival was censored for patient death with function. Univariate analyses for graft survival were performed using the Kaplan-Meier method. Cox proportional hazards model was used for the association between donor and recipient data, and graft survival (multivariate analyses). In recipients with graft loss, CrCl was set to the minimal value of 10 mL/min. P values ⬍.05 were considered significant in two-tailed tests. All analysis was done using SPSS software, version 11.5 (SPSS Inc, Chicago, Ill).
RESULTS
The study included 214 kidney biopsies (214 recipients) from 129 donors. Mean donor age was 42.2 ⫾ 16.6 years (range, 2 to 79) with 74.2% men. Overall mean recipient age was 47.7 ⫾ 13.5 years (range, 19 to 71) with 64% men. The main causes of donor death were cardiac disease (36.2%), cerebrovascular disease (31.9%), and trauma (25.7%). There were 58 (45%) NHBD and 71 (55%) brain dead donors. The mean number of glomeruli was 80.9 ⫾ 42.0 (range, 7 to 222) with 96.9% of biopsies containing ⬎20 glomeruli. GS was 6.8 ⫾ 7.8% (range, 0 to 33). In 18 (8.4%) recipients the graft had ⬎20% GS. IF was evident in 158 (73.8%) biopsies. Any grade of TA, AS or AH was identified in 139 (65%), 123 (57.5%), or 129 (60.3%) cases, respectively. A high correlation was observed among the histologic parameters, demonstrating them to be mutually interdependent (r range, 0.51 to 0.97; P ⬍ .001). All histologic variables were significantly associated with donor age (r range, 0.50 to 0.57; P ⬍ .001). In donors whose death was due to cerebrovascular or cardiac diseases, there were more severe histologic alterations. Histologic Parameters and Graft Outcome
Fifteen recipients lost their grafts in the first 3 posttransplant years. GS ⬎20% (log-rank, 7.9; P ⫽ .005), IF moderate and severe (Banff grade 2 to 3; log-rank, 10.4; P ⫽ .001), TA moderate or severe (Banff grade 2 to 3; log-rank, 8.1; P ⫽ .004), and AH moderate or severe (Banff grade 2 to 3; log-rank 14.2; P ⬍ .001) were significantly associated with graft loss. The significant association persisted after adjust-
1369 Table 1. Graft survival: Multivariate Analyses (Cox) Variables
P
Exp(B)
95% CI
Glomerulosclerosis (⬎20%) Interstitial fibrosis (grade 2–3) Tubular atrophy (grade 2–3) Arteriosclerosis (grade 2–3) Arteriolar hyalinosis (grade 2–3) Donor age* Acute rejection episodes (n) HLA mismatches (0–6) CMV disease Cold ischemia time Panel reactive antibody
.008 .003 .011 .051 .000 .384 .183 .409 .293 .129 .162
4.49 4.98 4.12 2.11 2.77 1.02 1.67 .83 .52 .90 1.01
1.4–13.7 1.7–14.3 1.4–12.2 .9–5.9 1.6–4.8 .9–1.7 .8–3.6 .5–1.3 .2–1.8 .8–1.0 1.0–1.03
*Although donor age did not show statistical independent association with graft loss, it was the variable more associated with graft function in the first 3 posttransplant years.
ing for donor age. AS was not independently associated with graft outcome. In the multivariate analysis for graft outcome, GS, IF, TA, and AH were independently associated with graft loss, controlling for donor age, AREs, HLA mismatches, CMV disease, cold ischemia time, and PRA (Table 1). Histologic parameters correlated with CrCl at 1, 2, and 3 posttransplant years, but, controlling for donor age only, IF and AH continue to show significant associations. In the multivariate analysis for graft function (CrCl at 1, 2, and 3 posttransplant years), no independent association was found between histologic variables and function, but increased donor age, number of acute rejection episodes, and CMV disease were independently associated with worse graft function. Graft outcome was not statistically different among NHBD and brain dead donors. DISCUSSION
Previous works have suggested that worse graft survival and function is associated with kidneys showing redused functional renal mass, chronic tubulointerstitial alterations, and arterial and/or arteriolar chronic lesions.1– 4,7 Kidney donor biopsy is more frequently performed for organs from older donors. Donor age is a known risk factor for allograft outcome, but it is not clear whether the histologic study provides additional prognostic information. Some studies have reported a poor correlation between histologic parameters and graft outcome,5,6 while others report an association with donor age.1– 4,8 In the present work we searched for an independent association between histologic parameters and graft survival/function. The initially formulated question was: is age the main donor risk factor, and do morphologic parameters provide additional predictor factors? Because it is clear that severely damaged kidneys are not transplanted, the question is the impact of mild or moderate chronic alterations. Our work has shown a modest association between histologic lesions and graft outcomes; donor age was a better predictor. The association between morphologic parameters and graft survival was clearer than the associa-
1370
tion with graft function. The explanation for this result is not evident. It is possible that more extensive chronic changes impact on outcome earlier. Upon multivariate analysis for survival, there was a low number of events: 15. Thus, there was a low reliability of results in a multivariate analysis. IF and AH were the microscopic variables with more significant P values for association with graft survival. In other works, morphologic parameters variably correlate with posttransplant graft outcome; the differences in results can be due to the methodology of the studies, the number of patients, the categorization of variables, a limitation to biopsies with only mild histologic changes, and lack of control for confounding factors. IF, GS, or AH have variably been identified as critical parameters in several studies. In the present work we found that IF and AH were similar to or better than GS as predictors of graft outcome. Thus, it is important to evaluate vascular and interstitial parameters as precisely as GS in a donor renal biopsy. Although our work showed donor age to be a better donor predictor, this statement did not mean that the donor biopsy is not important. Several cases were discarded because severe histologic changes, and the microscopic studies were decisive. Furthermore knowledge of the baseline morphologic status of an organ is important for subsequent examination and diagnosis in graft biopsies. In conclusion, in the present study, donor age was a better variable to predict graft outcome. Histologic param-
ARIAS, BLANCO, SANCHEZ-FRUCTUOSO ET AL
eters showed a modest association with graft function, but microscopic evaluation provided important information in the decision whether to transplant a kidney.
REFERENCES 1. Escofet X, Osman H, Griffiths DFR, et al: The presence of glomerular sclerosis at time zero has a significant impact on function after cadaveric renal transplantation. Transplantation 75:344, 2003 2. Randhawa PS, Minervini MI, Lombardero M, et al: Biopsy or marginal donor kidneys: Correlation of histologic findings with graft dysfunction. Transplantation 69:1352, 2000 3. Leunissen KM, Bosman FT, Nieman FH, et al: Amplification of the nephrotoxic effect of cyclosporine by preexistent chronic histological lesions in the kidney. Transplantation 48:590, 1989 4. Seron D, Carrera M, Grino JM, et al: Relationship between donor renal interstitial surface and post-transplant function. Nephrol Dial Transplant 8:539, 1993 5. Ratner LE, Joseph V, Zibari G, et al: Transplantation from hypertensive cadaveric donbors. Transplant Proc 27:989, 1995 6. Nyberg G, Hedman L, Blohme I, et al: Morphologic findings in baseline kidney biopsies from living related donors. Transplant Proc 24:355, 1979 7. Karpinski J, Lajoie G, Cattran D, et al: Outcome of kidney transplantation from high-risk donors is determined by both structure and function. Transplantation 67:1162, 1999 8. Pokorná E, Vitko S, Chadimova M, et al: Proportion of glomerulosclerosis in procurement wedge renal biopsy cannot alone discriminate for acceptance of marginal donors. Transplantation 69:36, 2000