Proteinuria After Kidney Transplantation, Relationship to Allograft Histology and Survival

Authors


* Corresponding author: Fernando G. Cosio, Cosio.Fernando@mayo.edu

Abstract

Proteinuria is associated with reduced kidney allograft survival. Herein we assessed the association between proteinuria, graft histology and survival. The cohort included 613 kidney allograft recipients who had proteinuria (measured) and surveillance biopsies at 1-year posttransplant. Proteinuria >150 mg/day was detected in 276 patients (45%) and in 182 of these, proteinuria was below 500. In >84% of patients even low levels of proteinuria were associated with albuminuria. Proteinuria was associated with the presence of graft glomerular pathology and the use of sirolimus. Eighty percent of patients with proteinuria >1500 mg/day had glomerular pathology on biopsy. However, lower levels of proteinuria were not associated with specific pathologies at 1 year. Compared to no sirolimus, sirolimus use was associated with higher prevalence of proteinuria (40% vs. 76%, p < 0.0001) and higher protein excretion (378 + 997 vs. 955 + 1986 mg/day, p < 0.0001). Proteinuria was associated with reduced graft survival (HR = 1.40, p = 0.001) independent of other risk factors including, glomerular pathology, graft function, recipient age and acute rejection. The predominant pathology in lost allografts (n = 57) was glomerular, particularly in patients with 1-year proteinuria >500. Thus, proteinuria, usually at low levels (<500 mg/day), is present in 45% of recipients at 1 year. However, and even low levels of proteinuria relate to poor graft survival. Proteinuria and glomerular pathology relate independently to survival.

Introduction

Kidney allograft survival, beyond the first year posttransplant, generally relates to variables that can be measured during the earlier posttransplant period. These predictive variables include, for example, the demographic characteristics of the donor and of the recipient; noxious events that might have occurred during the first year posttransplant, such as acute rejection; and allograft function and blood pressure measured during or at 1-year posttransplant. More recent studies have shown that certain histologic features of the allograft, independently from clinical variables, also relate to graft survival. For example, the presence of moderate or severe interstitial fibrosis (1,2), evidence of persistent inflammation and transplant glomerulopathy (3) relate to reduced allograft survival

In addition to the variables listed above, several studies have shown that proteinuria, even at low levels, portends poor graft survival (4,5). The nature of that relationship is unclear, in part because no previous study has evaluated systematically the histology of allografts with proteinuria. This relationship is likely due in part to the fact that proteinuria is a marker of recurrent or de novo glomerular diseases that are associated with poor prognosis (3,6,7). In addition, we hypothesized that proteinuria has negative prognostic implications for the allograft even in patients without readily recognizable glomerular disease either because focal glomerular lesions are missed in the biopsy or because proteinuria may relate in part to the severity of tubular dysfunction (8). To test this hypothesis we assessed allograft pathology and proteinuria 1-year posttransplantation in a large cohort of kidney recipients. We analyzed: (1) the relationship between proteinuria and allograft pathology from surveillance biopsies; (2) the relationships between proteinuria and graft survival in the presence or absence of glomerular pathology; and 3) the pathology of allografts that were lost during follow-up. Consistent with previous studies, these results showed that proteinuria, usually at low levels, is quite common following kidney transplantation and that even low levels of proteinuria (150–500 mg/day) relate to poor graft survival (4,5). As expected, patients with higher levels of proteinuria (>1500 mg/day) frequently had glomerular pathology. However, biopsies from patients with lower levels of proteinuria frequently have no visible glomerular pathology. Still in these patients proteinuria portends reduced graft survival. Examination of biopsies from kidneys lost during follow-up showed a somewhat unexpected high incidence of glomerular pathology, particularly in recipients who had proteinuria at 1 year.

Materials and Methods

Patient population

The study cohort includes adult kidney allograft recipients, transplanted at our institution between January 1998 and December 2004. Patients who were not followed at our institution and recipients of pancreas transplants before, after or at the time of the kidney transplant were excluded. Recipients of pancreas allografts were excluded because, at our institution they were historically bladder drained, and this would have confounded the assessment of proteinuria. Among this initial cohort of 960 patients, we selected 613 recipients who, 1-year posttransplant, had both a surveillance allograft biopsy and proteinuria measured in a 24-h urine collection. The surveillance biopsy and the 24-h urine collection are part of the standard first annual evaluation at our institution. Patients were excluded from these analyses if they had no surveillance biopsy (N = 46), no urine collection (n = 51) or neither (n = 250). Since 2004, measurement of urine albumin excretion has been added to the routine measurements done in the 24-h urine collections. Table 1 displays the demographic characteristics of the study population. Follow-up time for the cohort was 46 ± 21 (range 13–100) months.

Table 1.  Patient characteristics
ParameterValue
  1. 1Percentage of patients receiving no dialysis prior to the transplant.

  2. 2Borderline rejection is included.

Recipient age (years)51.1 ± 14.2
Donor age (years)41.4 ± 13.2
Recipient gender (% M:F)55.3:44.7
Caucasian race94.1%
Primary renal diagnosis
 Glomerulonephritis194 (32%) 
 Diabetes87 (14%)
 Polycystic KD80 (13%)
 Retransplant55 (9%) 
 Other197 (32%) 
Preemptive transplant143.6%
First transplant89.2%
Donor type
 Living related48.1 %
 Living unrelated28.1 %
 Deceased donor23.8 %
Type of transplant
 Conventional 544 (88.7 %)
 Positive cross match 47 (7.7 %)
 ABO incompatible 22 (3.6 %)
Acute rejection during first year2 117 (19.1%)

Recipient and donor clinical and laboratory information was extracted from electronic databases and from the medical record. The institutional review board approved this study and the collection of data. Graft function was measured at frequent intervals by serum creatinine and estimated glomerular filtration rate using the abbreviated modification of diet in renal disease (MDRD) equation (9). In addition, nonradiolabeled iothalamate clearance (10) was used to measure glomerular filtration rate (GFR) at three weeks post transplantation (at the time of dismissal of the patient from the early posttransplant clinic), at 1 year and yearly thereafter.

Renal pathology

As part of the routine clinical care of patients transplanted at our institution, surveillance allograft biopsies are taken within 30 min of implantation of the graft (time zero) and at 4,12, 24 and 60 months posttransplant. All biopsies were evaluated by one of five dedicated renal pathologists. Light microscopic findings are scored using the Banff ‘97 classification (11). Immunofluorescence, including C4d staining and electron microscopy are done only if clinically indicated. The diagnosis of acute rejection is based on biopsy evidence (11). One-year protocol biopsies were examined and findings related to the level of proteinuria. In addition, the pathology of allografts lost during follow-up was examined to determine the relationship, if any, between proteinuria and ‘terminal’ allograft histology.

Immunosuppression protocols

Immunosuppression consisted of induction with antithymocyte globulin in 463 (75.5%) cases. A smaller number of patients, 76 (12.4%) received induction with antiCD25 antibodies; four patients received OKT3; two received alemtuzumab; and two were enrolled in a blinded trial of either placebo or anti-CD25 antibodies. Sixty-six patients (10.8%) did not receive induction. At 1-year posttransplant, maintenance immunosuppression consisted of prednisone, tacrolimus and mycophenolate mofetil in 479 patients (78%). Fifty-three (8.6%) patients were on cyclosporine and sirolimus was used in 80 (13.1%). Patients treated with sirolimus posed a particular problem because during the first year there were frequent conversions to and from sirolimus. Overall 89 (14.5%) patients were exposed to sirolimus at different times during the first year of transplant. For the purpose of defining the effect of sirolimus on proteinuria and graft histology, we considered a patient to be treated with sirolimus if they were taking this drug at the time of their 1 year evaluation visit (n = 80 patients, 13.1%). Among these patients, 40 were started on sirolimus at the time of transplant and remained on the drug throughout the first year. The remaining 40 patients were switched from tacrolimus to sirolimus sometime during the first year of transplant, most often due to reduced graft function and/or development of moderate fibrosis in a protocol biopsy.

Statistical analysis

Data are expressed as means and standard deviation unless otherwise stated. Numeric data were analyzed using Student's t-test and ANOVA when data were approximately normally distributed, or with nonparametric tests for heavily skewed data, such as proteinuria. Nominal data were analyzed using the chi-square test. The relationships between proteinuria and clinical parameters were tested as a continuous variable after logarithmic transformation of the urine protein values and then by multiple linear regression analyses. Graft survival, censored for patient death, was analyzed by Kaplan–Meier methods and univariate or multivariate Cox models. The later tests were also used to determine the relationships between clinical and histologic variables and the end point of interest.

Results

Incidence of posttransplant proteinuria

Table 2 displays levels of proteinuria 1 year after transplantation. As can be seen, 276 of the 613 patients (45%) had abnormal protein excretion (protein excretion >150 mg/day). In the majority of cases (182 of the 276, 66%) urine protein excretion was below 500 mg/day. To assess the possible contribution of proteinuria from native kidneys, we reanalyzed these data considering separately patients who received preemptive transplants (n = 267) and those who were on dialysis prior to the transplant (n = 346). As can be seen in Table 2, data from these two subsets of patients did not differ significantly.

Table 2.  Protein excretion 1 year after transplantation in the entire cohort (all patients), in patients transplanted preemptively and patients maintained on dialysis prior to the transplant (nonpreemptive transplants)
Urine protein (mg/day)All patientsPreemptive transplantsNon preemptive transplantsp
  1. 1Normal range.

  2. 2Values represent the number of patients and percentage of all of the patients with proteinuria within the indicated range.

  3. 3Chi-square comparing preemptive to nonpreemptive transplantation.

  4. 4Mann–Whitney comparing preemptive to nonpreemptive transplantation.

N613267346 
Mean (range)453 ± 1188 (2–11870)328 ± 941 (4–10588)550 ± 1341 (2–11870)0.1644
0–1501 337 (55%)2161 (60%)176 (51%) 
151–500182 (30%) 78 (29%)104 (30%) 
501–150050 (8%)17 (6%)33 (9%) 
1501–300027 (4%) 7 (3%)20 (6%)0.2453
>300017 (3%) 4 (1%)13 (4%) 

Urine albumin levels were measured at 1 year in 166 patients. Abnormal microalbuminuria (>30mg/day) was present in 19% of patients who had normal urine protein excretion (<150 mg/day). In contrast, 27 of 32 (84%) patients with proteinuria between 150 and 500 mg/day had albuminuria and all patients (100%) with proteinuria above 500 mg/day had abnormal albumin excretion in the urine.

Proteinuria and clinical variables

Higher prevalence of proteinuria related to several clinical variables including female donor, male recipient and history of acute rejection. However, these associations were generally weak and as shown in Table 3, quantitative differences in proteinuria among these groups were generally small. In addition, compared to patients receiving tacrolimus or cyclosporine, patients receiving sirolimus had significantly higher levels of protein in the urine (377 vs. 955 mg/day, p < 0.0001) and higher prevalence of proteinuria (40% vs. 76%, p < 0.0001) (Table 3). There were significant differences in proteinuria between patients who received sirolimus since the transplant and those who were switched from tacrolimus to sirolimus sometime during the first year (Table 3). Both groups of sirolimus treated patients had significantly more proteinuria than patients not on sirolimus. However, patients switched from tacrolimus to sirolimus had more proteinuria than those on sirolimus since the transplant. Among the 40 patients switched to sirolimus, 10 had proteinuria measured shortly before the switch. Of these 10 patients, proteinuria was: <150 mg/day in 2; between 150 and 500 mg/day in 4; and >500 mg/day in the remaining 4.

Table 3.  Relationship between proteinuria at 1 year and other clinical variables
VariableNUrine protein (mg/day)Proteinuria >150, mg/day (%)
  1. 1Mann–Whitney; 2chi-square; 3p = 0.001 and 4p < 0.0001 comparing no sirolimus to sirolimus; 5p = 0.014 lower than the ‘switched to sirolimus’ group.

Donor sex 
 Females326461 ± 104452
 Males287444 ± 133538
  p-value 0.00110.0072
Recipient sex 
 Females274347 ± 979 37
 Males339539 ± 132851
  p-value <0.000110.012
Acute rejection 
 No496404 ± 112742
 Yes117660 ± 140657
  p-value 0.00110.0052
Immunosuppression 
 Sirolimus 80955 ± 198676
 No sirolimus533377 ± 997 40
  p-value <0.00011<0.00012
 Sirolimus since transplant 40   505 + 10003.572
 Switched to sirolimus 40 1504 + 2563480
 No sirolimus533377 + 99740
  p-value <0.00011<0.00012

Proteinuria at 1-year posttransplant did not relate significantly to the primary renal diagnosis. Specifically, proteinuria in patients with the pretransplant renal diagnosis of glomerulonephritis was not significantly higher than, for example, in patients with polycystic kidney disease (503 ± 1412 vs. 291 ± 993 mg/day, p = 0.9 Mann–Whitney). In addition, proteinuria at 1-year posttransplant did not differ statistically according to donor age, recipient age, donor or recipient race, history of diabetes, donor type (living vs. deceased), panel reactive antibodies (PRA) or HLA matching (data not shown).

Increasing levels of proteinuria related to other clinical parameters measured at 1 year, including: (1) increasing levels of serum creatinine (r = 0.326, p < 0.0001, linear regression); (2) decreasing GFR levels (r =− 0.181, p = 0.016); and (3) increasing systolic (r = 0.219, p = 0.018) and diastolic (r = 0.117, p = 0.048) blood pressure.

Proteinuria and allograft histology

Surveillance biopsies obtained 1 year after the transplant were classified into the following groups (Table 4): (1) normal histology (n = 186, 30%); (2) acute rejection (n = 39, 6%); (3) allograft interstitial fibrosis and atrophy without any evidence of glomerular disease (n = 273, 45%); (4) glomerular disease of any type (n = 56, 9%); and (5) other pathologies (n = 59, 10%). The group of biopsies with interstitial fibrosis and tubular atrophy excluded cases where glomerular pathology, aside from global glomerulosclerosis, was demonstrated. The glomerular group included all biopsies with recurrent disease (10 of 613, 1.6%); transplant glomerulopathy (23 of 613, 3.8%) or de novo glomerular pathology (23 of 613, 3.8%) which most commonly was focal segmental glomerulosclerosis (FSGS) (17 of 613, 2.8 %). It should be noted that the ‘other’ group of biopsies include biopsies with histologic abnormalities but sometimes with uncertain diagnosis (see list of pathologies in Table 4). For example, allografts with arteriolar hyalinosis are included here recognizing that this anomaly may or may not be due to calcineurin toxicity. This classification was based principally on light microscopic finding because immunofluorescence and electron microscopic studies were not obtained in all biopsies.

Table 4.  Urine protein levels at 1-year posttransplant according to the histology of the 1-year surveillance biopsy
Histologic variantN (%)Proteinuria (mg/day), mean ± SD (range)
  1. 1p < 0.0001 Kruskall–Wallis indicating that the proteinuria in the glomerular group was significantly higher than the other four groups.

  2. 2This group includes the following pathologies: BK nephropathy (n = 19); insufficient (n = 2); arteriolar hyalinosis (n = 5); arteriosclerosis of donor origin (n = 22); recurrent sickle cell nephropathy (n = 1); interstitial nephritis (n = 8) and acute tubular necrosis (n = 2).

Normal186 (30.3)207 ± 534 (4–6091)
Acute rejection39 (6.4)262 ± 389 (8–1986)
Fibrosis/atrophy273 (44.5)229 ± 289 (2–1931)
Glomerular pathology56 (9.1) 2716 ± 2889 (33–11870)1
Other259 (9.6) 239 ± 369 (24–2738)

Table 4 and Figure 1 display levels of urine protein excretion in patients classified according to their allograft histology. Proteinuria was significantly higher in patients with glomerular pathology than in any other histologic group (p < 0.0001) and levels of proteinuria greater than 1500 mg/day were highly associated with the presence of glomerular disease (Figure 1). Thus, 35 of 44 patients (80%) with proteinuria >1500 mg/day had demonstrable glomerular pathology in the biopsy. In contrast, lower levels of proteinuria did not differentiate the various histologic groups. As can be seen in Figure 1, allograft fibrosis and atrophy is rarely associated with proteinuria >1500 mg/day (3 of 273 cases, 1%) but lower levels of proteinuria were present in 45% of these patients. There was no significant association between sirolimus use and particular histologic group (data not shown).

Figure 1.

Proteinuria in patients classified according to the allograft histology at 1 year. Open bars: proteinuria <150 mg/day (normal protein excretion); light gray bars: proteinuria between 151 and500 mg/day; striped bars: 500–1500 mg/day; and black bars: proteinuria >1500 mg/day. In this figure, patients with more than 1500 mg/day of proteinuria were considered as a single group to simplify the figure.

Proteinuria and graft survival

During a follow-up period of 57 ± 21 (range 13.3–110) months, 38 (6.2%) patients expired and 57 (9.3%) grafts were lost not due to patient death. By univariate Cox regression, the following variables related to reduced graft survival (Table 5): (1) younger recipient age; (2) history of acute rejection; (3) reduced graft function at 1 year, measured as either serum creatinine, estimated GFR or measured GFR; (4) use of sirolimus versus no sirolimus; (5) increasing levels of proteinuria; and (6) the presence of glomerular disease in the 1-year protocol biopsy.

Table 5.  Variables significantly related to graft survival based on Cox regression
VariableUnivariate HR (CI); p-valueMultivariate2 HR (CI); p-value
  1. 1Hazard ratio calculated per gram per day of proteinuria.

  2. 2Included are only those patients with normal histology, fibrosis-atrophy or glomerular pathology.

Recipient age0.98 (0.96–0.99)0.97 (0.94–0.99)
p = 0.014p = 0.027
Acute rejection2.58 (1.51–4.40)2.45 (1.32–4.51)
p < 0.0001p = 0.004
Estimated GFR0.94 (0.92–0.96)0.95 (0.93–0.98)
p < 0.0001p < 0.0001
Sirolimus2.64 (1.5–4.7) 
P = 0.001p = 0.22
Proteinuria11.39 (1.28–1.52)1.27 (1.13–1.43)
p < 0.0001p < 0.0001
Allograft histology22.19 (1.87–4.53)1.58 (1.02–2.47)
p < 0.0001p = 0.04

Figure 2A displays the relationship between sirolimus use and allograft survival. As can be seen, there were no significant differences in graft survival between patients not on sirolimus and those on sirolimus since the time of transplant. In contrast, the survival of patients started on tacrolimus at the time of transplant and then converted to sirolimus for clinical reasons (including deteriorating kidney function and/or moderate to severe fibrosis in previous biopsies) had reduced survival compared to no sirolimus. Thus, the reduced survival of patients on sirolimus, shown here, could be explained by a preselection bias for the use of sirolimus in patients with deteriorating graft function and/or histology.

Figure 2.

Figure 2.

(A) Kaplan–Meier death censored graft survival in the following groups of patients: no sirolimus (……); sirolimus initiated at transplant (——); and patients switched from tacrolimus to sirolimus during the first year posttransplant (–––). Log rank p < 0.0001. (B) Kaplan–Meier death censored graft survival in patients classified according to the level of proteinuria at 1-year posttransplant: proteinuria <150 mg/day (——); proteinuria between 150 and 500 mg/day (+——+); proteinuria between 500 and 1500 (□——□); and proteinuria > 1500 mg/day (↓——↓). Log rank p < 0.0001. (C) Kaplan–Meier death censored graft survival in patients with normal histology (); fibrosis and atrophy (–––); or glomerular pathology (——) on the 1-year surveillance biopsy. Log rank p < 0.0001.

Figure 2.

Figure 2.

(A) Kaplan–Meier death censored graft survival in the following groups of patients: no sirolimus (……); sirolimus initiated at transplant (——); and patients switched from tacrolimus to sirolimus during the first year posttransplant (–––). Log rank p < 0.0001. (B) Kaplan–Meier death censored graft survival in patients classified according to the level of proteinuria at 1-year posttransplant: proteinuria <150 mg/day (——); proteinuria between 150 and 500 mg/day (+——+); proteinuria between 500 and 1500 (□——□); and proteinuria > 1500 mg/day (↓——↓). Log rank p < 0.0001. (C) Kaplan–Meier death censored graft survival in patients with normal histology (); fibrosis and atrophy (–––); or glomerular pathology (——) on the 1-year surveillance biopsy. Log rank p < 0.0001.

Figure 2.

Figure 2.

(A) Kaplan–Meier death censored graft survival in the following groups of patients: no sirolimus (……); sirolimus initiated at transplant (——); and patients switched from tacrolimus to sirolimus during the first year posttransplant (–––). Log rank p < 0.0001. (B) Kaplan–Meier death censored graft survival in patients classified according to the level of proteinuria at 1-year posttransplant: proteinuria <150 mg/day (——); proteinuria between 150 and 500 mg/day (+——+); proteinuria between 500 and 1500 (□——□); and proteinuria > 1500 mg/day (↓——↓). Log rank p < 0.0001. (C) Kaplan–Meier death censored graft survival in patients with normal histology (); fibrosis and atrophy (–––); or glomerular pathology (——) on the 1-year surveillance biopsy. Log rank p < 0.0001.

Figure 2B displays the relationships between proteinuria and graft survival. Increasing levels of proteinuria were associated with progressive reduction in graft survival. In this figure, all patients with proteinuria greater 1500 mg/day (n = 44) were considered as a single group. Quantitatively, compared to patients with normal protein excretion, urine protein levels between 151 and 500 mg/day were associated with a hazard ratio (HR) of 2.45 (1.2–5.01), p = 0.014; proteinuria between 500 and 1500: HR = 6.07 (2.7–13.8), p < 0.0001; and proteinuria greater than 1500: HR = 14.3 (5.75–36.1), p < 0.0001.

Figure 2C displays the relationship between histology and graft survival. Compared to other histologic groups, patients with glomerular disease had significantly reduced graft survival. In contrast, patients with fibrosis and atrophy, as a group, did not have significantly reduced graft survival compared to patients with normal biopsy findings.

All of the variables displayed in Table 5 were entered into a multivariate model which included only patients with normal histology, fibrosis/atrophy or glomerular disease (n = 515). Reduced graft survival was associated with the following independent variables: (1) younger recipient age; (2) history of acute rejection; (3) lower graft function; (4) proteinuria; and (5) glomerular disease on the surveillance biopsy. Thus, proteinuria correlates to reduced graft survival independently from the presence of glomerular pathology on the biopsy. To reassess this finding, we analyzed the relationship between proteinuria and graft survival in patients with fibrosis and atrophy. As can be seen in Figure 3, compared to patients with normal histology, in patients with fibrosis/atrophy levels of proteinuria above 500 mg/dL were associated with progressive reductions in survival. Quantitatively, the hazard ratio associated with proteinuria between 150 and 500 was 2.15 (0.68–6.8), p = 0.19 while the hazard ratio associated with higher levels of proteinuria was 5.11 (1.4–19.2), p = 0.016.

Figure 3.

Kaplan–Meier death censored graft survival in patients with normal histology (n = 186) (——) or with fibrosis and atrophy without proteinuria (□——□) or with 150–500 mg/day or proteinuria (——) or with more than 500 mg/day of proteinuria (… … ). Log rank p = 0.046.

Histology of lost allografts

At the end of follow-up, 57 allografts were lost not due to patient death. In 55 of these patients, follow-up biopsies were available prior to graft loss. Forty-four of the 57 patients (77%) who lost their allograft had proteinuria at 1 year while 13 (23%) did not. Table 6 displays the pathology of lost allografts: glomerular disease was the most common pathology in failing allografts (27 of 55 cases, 49%), including 17 cases (31%) of transplant glomerulopathy and 10 cases (18%) with recurrent disease. The second most common cause of graft loss was chronic allograft nephropathy of unspecified etiology (16 of the 55 cases, 29%). The incidence of glomerular disease was not significantly different in patients with proteinuria less than 500 mg/day and in those without proteinuria (not shown). In contrast, the percentage of grafts loss due to glomerular disease was significantly higher in patients who at 1 year had proteinuria greater than 500 mg/day than in those with lower levels of proteinuria (66% vs. 25%, p = 0.030).

Table 6.  Pathology of allografts lost during follow-up
DiagnosisAll patients (n = 55) (%)Proteinuria <500 mg/ day at 1 year (n = 31) (%)Proteinuria >500 mg/ day at 1 year (n = 24) (%)
  1. Table includes results of 55 of the 57 cases where biopsies were available. The remaining two cases did not have biopsies beyond 1-year posttransplant.

  2. 1Diagnoses included: three focal segmental glomerulosclerosis; three membranoproliferative glomerulonephritis; one IgA nephropathy; one recurrent anti-GBM disease and two diffuse proliferative glomerulonephritis.

  3. 2Diagnoses included: recurrent sickle cell nephropathy; chronic pyelonephritis; infiltrating lymphoma; and acute tubular necrosis.

Transplant glomerulopathy17 (31) 7 (23)10 (42)
Recurrent glomerulonephritis110 (18) 4 (13) 6 (25)
Chronic allograft nephropathy16 (29)12 (39) 4 (17)
Polyoma nephropathy 6 (11) 4 (13)2 (8)
Other24 (7)2 (6)2 (8)
Acute rejection2 (4)1 (3)1 (4)

Discussion

These analyses show that proteinuria, defined as a daily urinary protein excretion greater than the normal laboratory reference range, is present in approximately half of kidney allograft recipients 1 year after transplantation. In most cases, the urine protein excretion is low, often below 500 mg/day. However, even these low levels of proteinuria are associated with reduced graft survival and that association is independent of other prognostic variables such as graft function and acute rejection. These observations are in general agreement with previous studies (4). The primary purpose of this study was to assess the kidney pathology in transplant recipients with proteinuria. Guiding these studies were the following postulates, that proteinuria was often an indicator of allograft glomerular disease and also that, even in patients without detectable glomerular disease, proteinuria was associated with poor allograft prognosis. These results support, at least in part, both of these postulates. First, at 1 year 80% of patients with proteinuria greater than 1500 mg/day had histologic evidence of glomerular disease and second, in patients without detectable glomerular disease, such as patient with fibrosis and atrophy, proteinuria was associated with reduced allograft survival. These studies also shed some light on the nature of the association of proteinuria and graft prognosis and overall emphasize the importance of glomerular disease to graft survival, perhaps a previously underappreciated association. That is, examination of biopsies of allografts lost during follow-up showed that the majority of these patients (77%) had proteinuria at 1 year. Furthermore, glomerular pathology was the most common (presumed) cause of graft loss among patients in this cohort and this association was particularly striking in patients with history of proteinuria where 67% of lost allografts had either de novo or recurrent glomerular disease.

These observations are based on a transplant population that includes high proportions of Caucasians, living donor recipients and preemptive transplants. These characteristics may raise questions about the applicability of these findings to other transplant populations. However, several previous studies analyzing populations of varied characteristics (4,5) are in agreement with the relevance of proteinuria for allograft prognosis. Furthermore, the posttransplant histology of living donor (2) and of deceased donor allografts (12) is quite similar.

Glomerular diseases are relatively rare in kidney allografts. For example, in this study, 1-year surveillance allograft biopsies detected glomerular pathologies in 9% of allografts and in the majority of these the glomerular disease was transplant glomerulopathy. It should be noted that in these studies we separated these cases of glomerular pathology from the larger group of patients with chronic allograft nephropathy (now called fibrosis and atrophy [13]). Consistent with this approach, in the most recent iteration of the Banff criteria for interpretation of kidney biopsies (13), biopsies with transplant glomerulopathy are separated from the larger group of patients with fibrosis and atrophy. In contrast, de novo FSGS is still considered as part of the changes that may be associated with chronic allograft nephropathy (11). In contrast, the results of this study and others (6) emphasize that chronic allograft nephropathy is rarely associated with de novo FSGS and that this form of glomerular disease is associated with proteinuria and particularly poor graft survival. The poor prognosis of transplant glomerulopathy (3,14), recurrent glomerular disease (7) and de novo FSGS (6) explains why glomerular pathology is overrepresented in lost allografts. The contribution of glomerular disease to graft prognosis is notable yet poorly understood. There are interesting parallels to be noted between these observations in renal allografts and in other diseases primarily of the renal tubules and interstitium, such as reflux nephropathy. In that conditions, many years ago, it was also observed that the proteinuria at any level was associated with the development of glomerular disease and poor prognosis (15).

Previous studies indicated that even low levels of proteinuria (<500 mg/day) are predictive of reduced kidney graft survival (4). However, the reason for this association is only partially clarified by these results. One unifying hypothesis, suggested by these data, is that even at low levels, proteinuria may be indicative of glomerular pathology that carries a particularly poor prognosis. However, these analyses do not provide direct support for this hypothesis as allografts with this level of proteinuria do not have a high incidence of glomerular disease at 1 year or at the time of graft loss. We should first consider that proteinuria maybe mitigated by the hemodynamic effects of calcineurin inhibitors. Thus, ‘low levels’ of proteinuria may represent partially suppressed proteinuria (16). Low levels of proteinuria may be due to several additional mechanisms, including: residual proteinuria from native kidneys (17,18); glomerular hyperfiltration; effects of sirolimus (16,19–21); and/or tubular defects in absorption of albumin (8,22). In recent studies, we and others showed that even very high levels of proteinuria originating from native kidneys declines rapidly during the first 3–4 weeks posttransplant and continue to decline during the first year (17,18). However, very low levels of proteinuria, generally <500 mg/day, may remain in these patients. From previous studies we learned however that proteinuria greater than 1500 mg/day at 1 year and/or an increase in proteinuria from early posttransplant to 1 year of 500 mg/day or more should not be attributed to native kidney proteinuria because it is almost always associated with de novo allograft glomerular pathology (17). These analyses suggest, albeit indirectly, that glomerular hypertension/hyperfiltration may also play a role in posttransplant proteinuria because this anomaly was more common in female donor kidneys (generally smaller) and male recipients (generally larger body size).

Only a fraction of patients in this cohort had albuminuria measured directly. However, it is of interest to note that albuminuria was present in 20% of nonproteinuric patients and that most patients (84%) with proteinuria <500 mg/day and 100% of patient with higher levels of proteinuria had microalbuminuria. This finding is indeed suggestive of defects in glomerular permeability even in patients with low levels of proteinuria (23). Of interest, recent studies by Halimi et al. (24) linked albuminuria, even in the absence of proteinuria, to reduced allograft and patient survival. The latter association was not confirmed in this study where proteinuria did not relate statistically with patient survival.

The relationship between proteinuria and the fibrosis/atrophy pathology is partially clarified by these results. First, proteinuria occurs in approximately half of the patients with fibrosis and atrophy but it is rarely above 500 mg/day. Thus, higher levels of proteinuria should not be attributed to fibrosis/atrophy alone and most likely indicates the presence of glomerular pathology. These studies also emphasize that in these patients any level of proteinuria should alert the clinician of poor allograft survival prognosis. The reason for this latter association is unclear. First, examination of the pathology of lost graft indicates that in some of these allografts glomerular pathology will become evident during follow-up. Previous studies suggest that the proteinuria is not indicative of the severity of fibrosis/atrophy (25). However, as shown here, proteinuria relates to other adverse indicators of graft survival including elevated blood pressure and lower allograft function. Thus, proteinuria, aside from being a sensitive indicator of glomerular pathology, is also likely a sensitive indicator of a ‘troubled’ allograft. Finally, we should also consider the potential pathogenic effect of proteinuria itself on allograft prognosis. This hypothesis is well developed in native kidney diseases (26), although it is usually associated with higher levels of proteinuria. It has been suggested that reductions in proteinuria with angiotensin II inhibition is associated with improved prognosis in these patients (4).

This study confirms previous observations that the use of sirolimus is associated with an increased incidence of de novo proteinuria (16,19–21). However, we noted here that the proteinuria is more marked in patients switched from tacrolimus to sirolimus, generally due to allograft dysfunction, than in patients on sirolimus since the transplant. Based on similar observations it was suggested previously that the proteinuria that develops in patients switched to sirolimus is at least in part due to the loss of the antiproteinuric effect of the calcineurin inhibitor rather than a direct proteinuric effect of sirolimus (16). In these studies, we also noted an association between the use of sirolimus and reduced allograft survival. However, these results are most likely indicative of a selection bias and not necessarily indicative of a noxious effect of sirolimus on allografts. That is, the negative association between sirolimus and graft survival was noted only in patients switched from tacrolimus to sirolimus and not in patients on sirolimus since the transplant. In this and other cohorts, patients are generally switched to sirolimus because they have graft dysfunction already and thus, predictably they had poor graft survival. In agreement with previous studies (21), de novo immunosuppression with sirolimus is associated with proteinuria, a finding of concern particularly in view of the poor renal prognosis of proteinuric patients treated with sirolimus (27,28). Based on these data, it is prudent to recommend that sirolimus should not be started in patients with preexisting proteinuria (28) and that if de novo proteinuria, particularly if >500 mg/day, develops in a patient treated with this drug, the clinician should consider discontinuation of the sirolimus.

Multiple lines of evidence support the hypothesis that death censored kidney allograft survival is, in large part, predictable by clinical and biochemical markers measured during the first year posttransplant. In previous studies, we and others (3,29–32) have added histologic markers at 1 year to the list of independent variables that can be used to predict the life expectancy of a kidney allograft. This study shows for the first time that the proteinuria and graft histology are statistically independent predictors of graft survival. We propose that in part, the relationship between proteinuria and survival is due to a previously underestimated incidence of allograft glomerular pathology.

Acknowledgments

This work was supported by grants from the Mayo Clinic Transplant Center and Division of Nephrology and Hypertension. We thank the kidney-pancreas transplant coordinators for their dedication to the care of transplant recipients and their help in the collection of data from these patients. We also thank Ms. Cynthia Handberg for her excellent secretarial assistance.

Ancillary