Renal transplantation is the best therapeutic approach for patients with end-stage renal failure, contributing to a better quality of life and lower mortality when compared with dialysis (1). Currently, 1-year graft survival for recipients of cadaveric allograft is approaching 90%, but its half-life survival rate is still approximately 13 years (2). Therefore, most grafts will be lost after the first year because of death of the functioning graft and chronic allograft nephropathy (CAN). CAN is generally suspected when patients with previously stable renal function present a slow and progressive deterioration of renal-allograft function, commonly 3 or more months after transplantation (3). CAN designates an entity that has many indistinguishable features of chronic cyclosporine nephropathy. The diagnosis is confirmed by renal biopsy, in which the typical features are interstitial fibrosis and tubular atrophy with or without vascular intimal hyperplasia (4).
Although there are many clinical conditions associated with an increased risk of CAN development, there is no accurate clinical or laboratory test that can identify (among renal-transplant patients with good and stable graft function) those who present an enhanced risk of developing it. Thus, when the diagnosis of CAN is made, an important deterioration of renal-allograft function has already taken place, making any attempt to halt this process less successful.
Increased urinary levels of low molecular-weight proteins, such as β2 microglobulin and retinol binding protein (RBP), are a hallmark of proximal tubular dysfunction. The measurement of RBP in urine offers advantages over other proteins because its production is relatively constant; there is no known clinical situation in which overproduction could lead to abnormal urinary concentrations, and it is very stable in the whole range of urinary pH (5). Because CAN affects primarily the tubulointerstitial compartment, we reasoned that proximal tubular dysfunction, assessed by urinary levels of RBP (uRBP) detection, could be presented early in the course of this entity.
We have previously reported that heart-transplant recipients with high levels of uRBP had an increased risk of developing chronic renal failure during a 5-year follow-up (6). In addition, preliminary data obtained from a pilot study with a distinct group of renal-transplant patients showed a deterioration of graft function in those with high uRBP during a 1-year follow-up (7).
Here, we report on the results of a 5-year prospective study with 92 renal-transplant patients with good and stable graft function, in which uRBP levels predicted CAN development and dictated graft survival.
Among 400 recipients of kidney allografts performed between 1988 and 1996 in the Transplantation Unit at Universidade Federal de São Paulo, we initially selected 192 outpatients with stable graft function. From this group, we prospectively enrolled 92 patients in this study between 1995 and 1996. To be included, patients were to have stable and good renal function, namely, had at least three consecutive measurements of serum creatinine 1.5 mg/dL or less, proteinuria below 1.00 g/L at urinary sediment analysis, and absence of infection, graft rejection, or any unfavorable clinical conditions, with at least 3 months of posttransplant follow-up. Eighty-five (92.3%) patients were on cyclosporine-based immunosuppression, most of whom were on triple therapy with azathioprine and steroids, and seven (7.4%) of whom were receiving only azathioprine and steroids. All patients included in this study had normal serum hepatic enzymes levels. At the beginning of the study, the immunosuppression therapy was already tapered to maintenance doses.
Between 1995 and 1996, urinary samples were collected at least three different times from each patient (mean of 3.4 samples for patient) for RBP determination. The median posttransplant time at the enrollment was 9 (range 3.1–110) months. Patients with levels constantly above the upper limit of the reference interval were classified as high uRBP, and those with steady normal levels as normal uRBP. The median posttransplant time of uRBP measurement was 9.02 (range 3.33–84.4) and 8.85 (range 3.07–109.6) months for both groups, respectively. The uRBP samples were collected during a mean interval time of 10.2±6.4 and 9.0±5.6 months in high and normal uRBP groups, respectively (P=0.321). Demographic characteristics of patients distributed in both groups are shown in Table 1.
Delayed graft function was defined as the requirement of dialysis during the first week after transplant in absence of rejection or technical problems. Early acute rejection was assumed in patients with clinical suspicion who had had biopsy-proven diagnosis or had received methylprednisolone therapy until the third month of transplantation. After this period, it was presumed that patients were suffering from late acute rejection. Systemic arterial hypertension was considered when patients were using at least one antihypertension drug therapy. Diabetes mellitus was considered to be present if the patient was using any hypoglycemic drug at the time of urine sample collection.
CAN development was suspected when deterioration of renal function, with or without proteinuria or hypertension, was presented in a previously stable patient in at least two sequential visits 6 months posttransplantation. Graft biopsies were performed in all clinically suspected cases of CAN. A histopathologic criterion for CAN diagnosis was the presence of interstitial fibrosis and tubular atrophy with or without vascular intimal hyperplasia. A total of 28 graft biopsies were performed to confirm the presence of CAN.
RBP was determined by an immunoenzymometric assay, and the upper limit of normal was considered as being 0.400 mg/L (8). Urine samples were collected and frozen until the RBP determination. No conservative or special precautions were considered to be necessary because RBP is stable in urine. Ninety-six well plates (Nunc A/S, Roskilde, Denmark) were coated with 100 μL of a solution containing monoclonal antibody, 10 mg/L, in coating buffer. After an overnight incubation at 4°C, the wells were washed three times, and 100 μL of the samples were added. After being incubated for 2 hours at 37°C, the wells were washed and the biotinylated antibody added. After 1 hour of incubation, streptavidin-peroxidase and the color reagent (Amersham International, Little Chalfont, Buckinghamshire, UK) were added in sequence. Absorbances were read at 490 nm in a plate reader (Model EL 311; Bio-Tek Instruments, Winooski, VT). Concentrations in the samples were calculated by comparison with the standard curve, prepared by using nonlinear regression, usually as a third-degree polynomial. For statistical analyses, the level of uRBP was categorized as being high when above 0.400 mg/L and as normal when equal to or below 0.400 mg/L. For some analyses, uRBP was further categorized in levels below 0.400 mg/L, between 0.401 and 1.00 mg/L, and above 1.00 mg/L.
Pretransplant demographic characteristics used for covariate-adjusted analyses included the dialysis therapy (hemodialysis vs. peritoneal dialysis), source of the graft (cadaveric vs. living donor), and the age of the donor (above or below 43 years). The posttransplant-related variables included the presence of proteinuria in isolated samples (above or below 0.2 g/L), delayed graft function, early and late acute rejection, and cold ischemia time (above or below 33 hours). We examined the proportional-hazard assumption by plotting the graft survival curves for each group of a covariate on a log-log scale. Because the curves appeared reasonably parallel, we regarded the model as appropriate.
The Kaplan-Meier method was used to estimate CAN-free graft and graft and patient survivals in normal and high uRBP groups. Graft loss was defined by the requirement of permanent dialysis after graft failure, and death of patients with functioning grafts was not considered an end-point event. We determined statistical significance by making log-rank comparisons of survival curves using two-sided P values. Values are presented as mean and standard deviation (SD) and when appropriate, as median and ranges. Parametric and nonparametric tests, chi-squared, and Fisher’s exact test were performed to compare demographic covariates between groups when appropriate. A P value of less than 0.05 was considered significant. We used Stata statistical software 5.0 for all statistical analyses (Stata Corporation, College Station, TX).
Proximal Tubular Dysfunction Incidence
Among 92 patients, 48 (52%) had high uRBP (1.435 mg/L, range 0.415–22.6), whereas 44 (48%) had normal levels (0.091 mg/L, range 0.021–0.339). Patients with high and normal uRBP had similar sex distribution, age of recipients and donors, and percentage of cadaveric grafts (Table 1). Both groups had similar median posttransplant time and serum creatinine values at the enrolment (1.3 vs. 1.2 mg/dL, P=NS, respectively) (Table 1). Incidence and duration of delayed graft function and acute rejection and cyclosporine trough levels were also comparable in both groups (Table 1). Fifty-six percent of patients with normal uRBP and 66% of those with high uRBP had hypertension (P=0.449). Among hypertensive patients, calcium antagonist and beta blockers were the most frequent association (60% and 65%, respectively). ACE inhibitors were used only in 0.04% and in 0.06% of the patients with high and normal uRBP levels, respectively.
During a median follow-up of 60.6 months, 23 (25%) patients were diagnosed with CAN by graft biopsy, 19 (82.6%) of whom had high uRBP levels at enrollment period. A high level of uRBP was associated with a 96.5% increased risk of CAN development (relative risk of 1.96, 95% confidence interval [CI] 1.4066–2.7466, P=0.0007), with a negative predictive value of 90.9%. All patients who developed CAN or lost their graft had, at the beginning of the study, similar levels of serum creatinine and cyclosporine trough levels to those who did not develop CAN or did not lose their graft (Fig. 1). However, they had a significantly higher uRBP (1.149 mg/L, range 0.043–22.6) when compared with those who did not develop CAN (0.186 mg/L, range 0.021–8.950) (P=0.001).
Figure 2A shows Kaplan-Meier estimation of 5-year CAN-free survival for grafts from patients with normal uRBP (89.9%) and high uRBP (57.5%) (P=0.0004). CAN was diagnosed 20.4 (range 7.2–56.6) months after RBP determination in patients with high uRBP and after 27.9 (range 22.6–58.4) months in patients with normal uRBP. When patients were stratified in three levels of uRBP (≤0.400, 0.401–1.00, and >1.00 mg/L), we observed that 5-year CAN-free graft survival proportionally deteriorated as uRBP levels increased (Fig. 2B).
During the follow-up period, 11 (11.9%) patients lost their grafts. All of them belonged to the high uRBP group, consequently having levels of uRBP (2.6 mg/L, range 0.59–22.60 mg/L) significantly higher than the remaining 81 (88.1%) patients (0.300 mg/L, range 0.021–9.75 mg/L) (P=0.001).
Kaplan-Meier estimation of graft survival at 5 years for patients with normal uRBP (100%) and high uRBP (70.7%) (P=0.0002) is shown in Figure 2C. As the data shows, graft survival correlated with the level of uRBP. Five-year graft survival rate was worse in patients with uRBP levels superior to 1.00 mg/L. Although the curves of intermediate RBP (0.401–1.0 mg/L) and high RBP (>1.00 mg/L) were not statistically different, we have shown that unlike the highest uRBP group, the first graft lost within the intermediate levels of uRBP group occurred only after 3 years of follow-up (Fig. 2D). Patient survival rate at 5 years was better in patients with normal uRBP (97.7%) than in those with high levels uRBP (85.8%), although it did not reach a statistical significance (P=0.054).
Given the increased risk of CAN development in patients with high levels of uRBP (96.5%), we further analyzed the influence of other covariates in its development by using a Cox proportional hazard model. A high level of uRBP was the most powerful predictor of biopsy-proven CAN occurrence when compared with pre- (relative risk of 6.183, P=0.001) and posttransplant (relative risk of 5.271, P=0.004) demographic variables. uRBP level above 0.400 mg/L was associated with a five to sixfold increase in the risk of CAN (Table 2). Considering other variables, we observed that although most patients who developed CAN had hypertension, there was no statistical correlation (P=0.264) because the same trend was observed in patients that did not develop CAN.
Presence of Isolated Proteinuria
Proteinuria in isolated samples was also a significant covariate in the development of CAN in these patients. Eleven (11.9%) patients had protein levels between 0.2 and 0.9 g/L, eight of whom had high levels of uRBP. It is worth emphasizing that 40 of 48 (83.3%) patients who had high uRBP levels had no proteinuria. We additionally analyzed whether high uRBP levels were also implicated in an increased risk of CAN development and poor graft survival in patients without proteinuria. The incidence of biopsy-proven CAN among 81 proteinuria-free patients was 19.8% (16 patients). Thirteen of those patients diagnosed with CAN (81%) had high uRBP levels and 3 (19%) normal levels. A high level of uRBP was still associated with 95.9% risk of developing CAN (relative risk of 1.959, 95% CI 1.34–2.83, P=0.0053). Figure 3A shows the Kaplan-Meier estimation of CAN-free graft survival at 5 years for proteinuria-free patients with high uRBP (64.0%) and normal uRBP (92.5%) (P=0.0034). Among proteinuria-free patients, only those with elevated levels of uRBP lost their graft, with a 5-year graft survival of 79.4% (P=0.0042) (Fig. 3B). Considering pre- and posttransplant covariates, Cox proportional hazard model analyses in this specific population showed that a high uRBP level was the unique factor associated with CAN development (relative risk of 5.336, 95% CI 1.454–19.580, P=0.012) and was linked with more than a fivefold increased risk of developing CAN (Table 3).
Despite all improvements in immunosuppressive therapy, human leukocyte antigen matching, prevention, and cautious management of many of CAN’s risk factors, CAN persists as the major cause of renal-allograft failure (3, 9). Currently, there is no established treatment for CAN, but data from experimental models suggest that, in early stages, the blockage or removal of the injury factor can halt or even reverse CAN (10). If this is also true for human renal transplantation, to delay or reverse CAN progression, we will need to diagnose it early. In this sense, a practical method for early identification becomes a sine qua noncondition for the clinical management of transplant recipients at a critical risk of developing CAN.
Assessment of allograft function of renal-transplant patients is routinely made by serum creatinine, urinary sediment analyses and, more recently, by routine biopsies (11, 12). Although, in many progressive renal diseases, the degree of tubulointerstitial lesion correlates with long-term organ-function prognosis (13, 14), renal tubular function is not usually evaluated in transplant patients. In addition, it has been shown that tubular proteinuria is an earlier marker of proximal tubular dysfunction, compared with urinary detection of phosphate, glucose, and amino acids (15).
In this study, we found an unexpected high incidence (52%) of proximal tubular dysfunction in renal-transplant patients. It is noteworthy that our patients had very good and stable allograft function at the beginning of the study, making it difficult to identify any possible sign of progressive graft disease by current laboratory methodology. A high level of uRBP was associated with a fivefold increased risk of CAN development and subsequent graft loss. Even considering only patients free of isolated proteinuria, the incidence of CAN and graft loss were still elevated in patients with high uRBP. Although we have not addressed the impact of other variables related to chronic graft dysfunction such as cigarette smoking, obesity, and hyperlipidemia, we conceived that they might have had, in some extension, an influence on the outcome of our population.
Furthermore, the time between uRBP assessment and CAN diagnosis was almost 2 years, which strongly suggests that some therapeutic maneuver can be tried before graft function deteriorates in these patients. Patients with normal and stable renal function who had high levels of uRBP had an increased risk for CAN development and consequently graft loss.
Cyclosporine nephrotoxicity is one of the major nonimmunologic risk factors for CAN development (16). We have recently reported that heart-transplant patients with elevated levels of uRBP had a 3.87 times higher risk of developing progressive renal failure caused by cyclosporine nephrotoxicity when compared with patients with normal uRBP (6). We considered that in some patients in our study, nephrotoxicity might have played a role in the progression of graft dysfunction. However, cyclosporine nephrotoxicity cannot be implicated in all cases because among seven patients receiving only azathioprine and prednisone, four had high uRBP, two of whom developed CAN during the follow-up. These four patients with high uRBP had steady higher levels at enrollment time, and one who developed CAN had the highest observed level (up to 19.3 mg/L), although we could not differentiate them from patients under calcineurin inhibitors that developed CAN. The remaining three patients who were under calcineurin–inhibitor-free therapy had normal levels of uRBP at enrolment (up to 0.256 mg/L) and maintained their renal function preserved during all the study period. Furthermore, cyclosporine trough levels were not significantly different between patients with high and normal uRBP and those who developed CAN and who did not. This suggests that drug exposure or tissue sensitivity were possibly more important than isolated drug trough levels.
In our view, there are two reasons for the high uRBP levels found in our very selected patient group: high uRBP levels could be a marker of proximal tubular dysfunction, as a signal of distress of proximal tubular cells, or a marker of definitive tubular damage, as a result of tubulointerstitial fibrosis.
Tubulointerstitial compartment integrity seems to be crucial to the later development of CAN (17, 18). Tubular membrane rupture correlates with an increased risk of CAN (19). In this sense, persistent insults might trigger inflammatory and immunologic tissue responses that ultimately could lead to fibrosis. Indeed, it has been demonstrated that human proximal tubular cells are capable of stimulating renal fibroblasts by producing and releasing platelet-derived growth factor and transforming growth factor-β (20).
CAN is presently one of the major causes of end-stage renal disease leading to retransplant in the United States. The worldwide tendency to use marginal kidneys (i.e., grafts from older donors or with a pronounced ischemic insult) could enhance the incidence of CAN. Any attempt to prevent or to manipulate patients with CAN should include the early detection of this entity. A reliable, noninvasive, and simple test that can be applied in clinical routine is imperative. uRBP monitoring might contribute to this approach by being a functional parameter of proximal tubular cells that are affected early in CAN development and have a strong power of predicting graft loss.
The authors thank Dr. Nelson Zoccoler, Karen Scott, and Ifeoma Offiah for helpful discussions and critical review of the manuscript.
1. Wolfe RA, Ashby VB, Milford EL, et al. Comparison of mortality in all patients on dialysis, patients on dialysis awaiting transplantation, and recipients of a first cadaveric transplant. N Engl J Med
1999; 341: 1725.
2. Hariharan S, Johnson CP, Bresnahan BA, et al. Improved graft survival
after renal transplantation in the United States, 1988 to 1996. N Engl J Med
2000; 342: 605.
3. Halloran PF, Melk A, Barth C. Rethinking chronic allograft nephropathy: the concept of accelerated senescence. J Am Soc Nephrol
1999; 10: 167.
4. Racusen LC, Solez K, Colvin RB, et al. The Banff 97 working classification of renal allograft pathology. Kidney Int
1999; 55: 713.
5. Bernard AM, Moreau D, Lauwerys R. Comparison of retinol-binding protein and beta 2-microglobulin determination in urine for the early detection of tubular proteinuria. Clin Chem Acta
1982; 126: 1.
6. Camara NO, Matos AC, Rodrigues DA, et al. Early detection of heart transplant patients with increased risk of cyclosporin nephrotoxicity. Lancet
2001; 357: 856.
7. Pacheco-Silva A, Nishida S, Souza Silva M, et al. Urinary retinol-binding protein (RBPur) as a prognostic test for graft function in stable renal transplanted patients [Abstract]. J Am Soc Nephrol
1994; 5: 1029.
8. Pereira AB, Nishida SK, Vieira JG, et al. Monoclonal antibody-based immunoenzymometric assays of retinol-binding protein. Clin Chem
1993; 39: 472.
9. Monaco AP, Burke JF, Ferguson RM, et al. Current thinking on chronic renal allograft rejection: issues, concerns, and recommendations from a 1997 roundtable discussion. Am J Kidney Dis
1999; 33: 150.
10. Tullius SG, Hancock WW, Heemann U, et al. Reversibility of chronic renal allograft rejection. Critical effect of time after transplantation suggests both host immune dependent and independent phases of progressive injury. Transplantation
1999; 58: 93.
11. Rush DN, Henry SF, Jeffery JR, et al. Histological findings in early routine biopsies of stable renal allograft recipients. Transplantation
1994; 57: 208.
12. Seron D, Moreso F, Bover J, et al. Early protocol renal allograft biopsies and graft outcome. Kidney Int
1997; 51: 310.
13. Sesso R, Santos AP, Nishida SK, et al. Prediction of steroid responsiveness in the idiopathic nephrotic syndrome using urinary retinol-binding protein and beta-2-microglobulin. Ann Intern Med
1992; 116: 905.
14. Wehrmann M, Bohle A, Held H, et al. Long-term prognosis of focal sclerosing glomerulonephritis. An analysis of 250 cases with particular regard to tubulointerstitial changes. Clin Nephrol
1990; 33: 115.
15. Piscator M. Early detection of tubular dysfunction
. Kidney Int Suppl
1991; 34: S15.
16. Halloran PF. Risk factors in renal transplantation and implications for immunosuppressive therapy. Transplant Proc
1999; 31: 351.
17. Isoniemi H, Taskinen E, Hayry P. Histological chronic allograft damage index accurately predicts chronic renal allograft rejection. Transplantation
1994; 58: 1195.
18. Nickerson P, Jeffery J, Gough J, et al. Identification of clinical and histopathologic risk factors for diminished renal function 2 years posttransplant. J Am Soc Nephrol
1998; 9: 482.
19. Bonsib SM A-ES, Ahmad I, Young SM, et al. Acute rejection-associated tubular basement membrane defects and chronic allograft nephropathy. Kidney Int
2000; 58: 2206.
20. Johnson DW, Saunders HJ, Baxter RC, et al. Paracrine stimulation of human renal fibroblasts by proximal tubule cells. Kidney Int
1998; 54: 747.