The Effect of Everolimus Versus Mycophenolate Upon Proteinuria Following Kidney Transplant and Relationship to Graft Outcomes


Corresponding author: Alexander C. Wiseman


Although mTOR inhibitor use has been associated with proteinuria in kidney transplant recipients, dose dependency and impact on allograft function are unknown. In a post hoc analysis, we compared rates of proteinuria 3 months posttransplant among everolimus (EVR) and mycophenolate (MPA) treatment arms and used a time-dependent model to correlate the risk of proteinuria to EVR trough levels up to 24 months posttransplant. eGFR and graft loss was compared by proteinuria status at 3 months. Of 833 randomized patients, 24%, 36% and 19% of lower exposure EVR (1.5 mg/day), higher exposure EVR (3.0 mg/day) and MPA-treated patients had proteinuria ≥ 300 mg/g Cr at 3 months, respectively. EVR 1.5 was not associated with an increase in risk of proteinuria (HR 1.20; p = 0.19) unlike EVR 3.0 (HR 1.84; p < 0.001) versus MPA. EVR trough levels >8 ng/mL were significantly associated with proteinuria compared to 3–8 ng/mL (HR 1.86; p < 0.001). Those patients with proteinuria at 3 months and those who developed proteinuria thereafter had lower eGFR and higher graft loss at 24 months, regardless of treatment arm. We identify a dose-dependent effect of EVR with the risk of proteinuria; however, its independent impact upon eGFR and graft survival at 2 years was not evident.


angiotensin converting enzyme inhibitor/angiotensin receptor blocker;


calcineurin inhibitor






estimated glomerular filtration rate




hazard ratio


myfortic, mycophenolate;


mammalian target of rapamycin




odds ratio


protein/creatinine ratio;


standard deviation;




Proteinuria is a common finding following kidney transplant, affecting nearly 20% of transplant recipients when stringent criteria (>1 g/day) are applied [1, 2], and nearly 30–40% of recipients when less stringent criteria (>300 mg/day) are applied [3, 4]. Although proteinuria is associated with diminished patient and kidney graft survival [5], many factors contribute to the formation of proteinuria, rendering a direct causal relationship between proteinuria and poor outcomes difficult. In addition, the disease processes that contribute to proteinuria vary by the time posttransplant at which it is identified, further confounding the ability to conclude that proteinuria independently contributes to graft dysfunction and/or death.

Among the causes of proteinuria following transplant, which include donor factors, recurrent primary renal disease, persistent native kidney disease, chronic immunologic injury and transplant glomerulopathy, includes the use of immunosuppressive agents from the mammalian target of rapamycin (mTOR) inhibitor class, sirolimus (SRL) and everolimus (EVR; Refs.  [6][8]). Potential mechanisms for mTOR-associated proteinuria include decreased vascular endothelial growth factor synthesis, and inhibition of key podocyte proteins that comprise the glomerular slit diaphragm, including nephrin [9, 10]. In particular, proteinuria has been reported following calcineurin inhibitor (CNI) withdrawal with mTOR introduction in both the early (1–6 months; Refs.  [11, 12]) and late (>6 months; Refs.  [13, 14]) posttransplant setting. In general, these studies do not show diminished renal function in the setting of early mTOR transition, but suggest worsening renal function when mTOR are introduced in later time points following transplant in the setting of preexisting glomerular injury.

Many questions still persist when examining the effect of proteinuria and mTOR use upon graft outcomes, such as the risk and progression of proteinuria with de novo use of mTOR, the relative increase in risk of proteinuria versus other de novo immunosuppressive regimens, whether proteinuria related to mTOR is a dose-dependent phenomenon in clinical practice and whether mTOR-related proteinuria is associated with decrements in glomerular filtration rate (GFR) or graft survival over time.

To address these issues, we examined proteinuria data collected over a 24-month period following transplant from a randomized controlled trial in which two dosing regimens of EVR plus low-dose cyclosporine (CSA) were compared to a standard CSA/mycophenolate regimen [15]. Our hypothesis was that EVR use would be associated with a dose-dependent risk of proteinuria following transplant, but that this enhanced risk of proteinuria was not associated with an increased risk of renal dysfunction or graft loss. We additionally sought to address whether proteinuria identified early following transplant (3 months following surgery) would be associated with a diminished GFR over time, and whether this was influenced by early mTOR use.

Materials and Methods

A post hoc analysis of proteinuria incidence and relationship to graft function and survival was performed using data from a multicenter, randomized prospective report comparing two regimens of EVR with low-dose CSA to a standard CSA/myfortic (MPA) control arm. The two dosing regimens of EVR were (i) 1.5 mg/day in two divided doses and targeted to trough concentration 3–8 ng/mL and (ii) 3.0 mg/day in two divided doses and targeted to trough concentrations of 6–12 ng/mL. The study design and clinical outcomes have been previously published [15]. In brief, the study's primary study endpoints were acute rejection, graft loss and death with secondary analyses of kidney function (by estimated GFR) over time. The pertinent findings included equivalent outcomes among the three groups for the primary and secondary endpoints while achieving a 60% reduction in CSA exposure in the EVR arms, with higher rates of side effects in the higher dose EVR arm. Within this study design included the provision for spot urine protein and urine creatinine measures at predefined time points (month 3, 6, 12, 18 and 24 following transplant). Proteinuria was defined as urine protein/creatinine (Pr/Cr) ratio > 300 mg/g when represented as a categorical variable, because of the relationship with renal disease progression in native kidney disease. An additional analysis of Pr/Cr ratio >1.5 g/g Cr was also performed, based upon recent data suggesting a strong relationship with transplant glomerular pathology on biopsy and graft outcomes [3]. When defining early proteinuria and determining progression of early proteinuria and relationship to GFR, graft loss and angiotensin converting enzyme inhibitor/angiotensin receptor blocker (ACEI/ARB) use, the month 3 time point was chosen to avoid confounding of proteinuria from native kidney disease, which in previous studies has been shown to dissipate at 4–8 weeks following transplant [16, 17].

Statistical methods

From the initial study cohort, patients in whom urine protein/creatinine ratios could be calculated at month 3 are included in the analysis. The baseline characteristics were compared between treatment groups for all patients with data at month 3, for patients with proteinuria at month 3 and for patients without proteinuria at month 3. An analysis of variance model with treatment as a factor was used to analyze the continuous variables. A chi-square test was used to compare the categorical variables. Similar analyses were done to compare the baseline between the group with proteinuria at month 3 and the group without proteinuria at month 3.

The rates of proteinuria were summarized by treatment from month 3 to month 24. A Wilcoxon rank sum test was used to compare the urine protein/creatinine ratios between each pair of treatment groups at each visit. A Cox proportional hazard regression model was used to estimate the risk of proteinuria. The model included proteinuria as a time-dependent variable with independent variables of treatment group, EVR target levels of 3–8 ng/mL (below, within and above target), diabetes and BMI.

An analysis of covariance model was used to compare mean GFR (MDRD) at various time points between treatment groups and to compare mean GFR (MDRD) between patients with and without proteinuria at month 3. Chi-square tests were used to compare the rates of graft loss of patients with and without proteinuria at month 3. The rates of graft loss between treatment groups were also compared. A similar analysis was also done comparing the rates of graft loss for patients with different levels of proteinuria (300–1500 mg/day vs. >1500 mg/day). Logistic regression modeling was also performed for the rates of graft loss. Odds ratios were calculated for treatment groups, and for proteinuria versus no proteinuria at month 3.


Of 833 patients, 277, 279 and 277 patients were initially randomized to EVR 1.5, EVR 3.0 and MPA arms, respectively (Table 1). For the assessment of changes in median proteinuria status over time and for time-dependent multivariate model, all samples available for all time points were utilized. The rationale for utilizing these data in this manner (an “on therapy” analysis) was to quantify the relationship of proteinuria status by immunouppressive agent (EVR vs. MPA) arm and to correlate proteinuria status to actual EVR trough level achieved (exposure). As shown in Figure 1, measured proteinuria was highest for the EVR 3.0 arm, followed by EVR 1.5 and MPA arms, with all differences between groups for all time periods achieving statistical significance (p < 0.05). The median levels of proteinuria did not increase over time but remained stable among all treatment arms in this “on therapy” analysis.

Table 1. Baseline characteristics
 EVR 1.5 (n = 277)EVR 3.0 (n = 279)MPA (n = 277)
Recipient characteristics   
 Age, years ± SD45.7 ± 12.745.3 ± 13.447.2 ± 12.7
 Male, n (%)176 (63.5)191 (68.5)189 (68.2)
 Caucasian, n (%)193 (69.7)180 (64.5)190 (68.8)
Primary disease, n (%)   
 Hypertension/nephrosclerosis50 (18.1)50 (18.1)45 (16.2)
 Glomerulonephritis/glomerular disease43 (15.5)43 (15.5)40 (14.4)
 Diabetes mellitus39 (14.1)39 (14.1)45 (16.2)
 Polycystic disease36 (13.0)36 (13.0)33 (11.9)
 Unknown34 (12.3)34 (12.3)39 (14.1)
 Other/missing74 (27.1)74 (27.1)74 (27.1)
0 HLA mismatches, n (%)10 (3.6)15 (5.4)15 (5.4)
Preemptive transplant46 (16.6)37 (13.3)46 (16.6)
BMI25.8 (5.1)25.8 (5.0)25.9 (4.7)
DGF27 (9.7%)29 (10.4%)26 (9.4%)
Cold ischemia time8.3 (8.7)8.2 (9.0)8.6 (9.4)
Acute rejection (with 3 month)25 (9.0%)20 (7.2%)32 (11.6%)
Most recent PRA ≥ 20%7 (2.6%)5 (1.8%)4 (1.5%)
Donor characteristics   
 Age, years ± SD41.4 ± 13.941.1 ± 13.041.8 ± 13.6
 Male, n (%)154 (55.6)139 (49.8)136 (49.1)
 Caucasian, n (%)193 (69.7)191 (68.5)197 (71.1)
 Living donor147 (53.0)151 (54.1)148 (53.5)
Figure 1.

Median urinary protein/creatinine ratio (mg/g) of patients receiving everolimus and lower exposure cyclosporine or mycophenolate and higher exposure cyclosporine, sampled at prespecified time points following transplant.

When examining proteinuria as a categorical value, at 3 months a total of 58/240 (24%), 87/242(36%) and 47/244(19%) of the EVR 1.5, 3.0 and MPA-treated patients had proteinuria ≥ 300 mg/g Cr, respectively, and all three groups had 3% incidence of proteinuria >1.5g/day (Table 2). At 24 months, those with proteinuria ≥ 300 mg/g Cr were 42/159 (26%), 45/153(29%) and 18/166(11%), respectively, with a higher fraction of these patients experiencing proteinuria >1.5g/day in the EVR arms (Table 2). Thus, although the median proteinuria values among treatment arms remained stable (Table 1), there were similar rates of proteinuria in the two EVR arms at 24 months in the “on therapy” condition and there was a numerical decrease in rates of proteinuria ≥300 mg/g Cr in the MPA arm particularly in the number of patients with proteinuria >1.5 g/day.

Table 2. Rates of proteinuria among the three treatment arms at months 3 and 24 (on therapy)
TimeProteinuria statusEVR 1.5EVR 3.0MPA
3 months≥ 300 mg/g58/240 (24%)87/242 (36%)47/244 (19%)
 300 to <1500 mg/g51/240 (21%)79/242 (33%)40/244 (16%)
 ≥1500 mg/g7/240 (3%)8/242 (3%)7/244 (3%)
 No proteinuria182/240 (76%)155/242 (64%)197/244 (81%)
24 months≥ 300 mg/g42/159 (26%)45/153 (29%)18/166 (11%)
 300 to <1500 mg/g33/159 (21%)35/153 (23%)17/166 (10%)
 ≥1500 mg/g9/159 (6%)10/153 (7%)1/166 (1%)
 No proteinuria117/159 (74%)108/153 (71%)148/166 (89%)

To determine if the risk of proteinuria in each EVR arm was dose dependent, a time-dependent multivariate analysis correlating proteinuria status with treatment arms and EVR trough level achieved was performed. As shown in Table 3, the likelihood of having proteinuria ≥300 mg/g was strongly associated with higher dose (3.0) and higher exposure (trough level > 8) EVR versus MPA (HR 1.84 and 1.86 respectively; p < 0.0001), but was not statistically higher for the EVR 1.5 arm or with EVR levels 3–8 ng/mL (HR 1.20; p = 0.19 and 1.24, p = 0.17, respectively). Proteinuria was also associated with diabetes status but was not associated with obesity (BMI > 30 kg/m2). Thus, a dose-dependent relationship of proteinuria with EVR exposure was evident, with a nonsignificant trend to higher proteinuria with EVR doses and trough goals that are currently supported in clinical use.

Table 3. Time dependent analysis results: HR for proteinuria > 300 mg/g at prespecified time points of assessment (months 3, 6, 9, 12, 18, 24)
 Hazard ratiop-Value95% Hazard ratio confidence limits
1.5 EVR versus MPA1.200.190.92–1.57
3.0 EVR versus MPA1.84<0.00011.40–2.40
EVR < 3 ng/mL versus 3–8 ng/mL1.240.17440.91–1.68
EVR > 8 ng/mL versus 3–8 ng/mL1.86<0.00011.37–2.52
Diabetes versus no1.70<0.00011.27–2.03
BMI > 30 versus <300.9930.500.97–1.01

Given this dose-dependent relationship of proteinuria and EVR exposure in the “on therapy” analysis, additional questions that may guide clinical practice include (i) whether proteinuria status was associated with an increase in study discontinuation (and thus confound interpretation for risk of reduced GFR or graft loss), (ii) whether proteinuria defined early after transplant is associated with poorer graft function and (iii) whether there is a more pronounced effect with one immunosuppressive agent (EVR) versus another (MPA) on progressive proteinuria or graft outcomes. To address these issues, patients with early proteinuria (3 months posttransplant) were selected for longitudinal analysis. Of the initial 833 participants, 726 had urinary protein/Cr measurements available at 3 months for post hoc longitudinal analysis (Table 4). The reasons for lack of inclusion in this analysis were, in the EVR 1.5 mg arm (n = 37 patients)-adverse event (23), unsatisfactory response [1], protocol violation [2], withdrew consent [4], administrative problem [3], death [1], graft loss [3]; in the EVR 3.0 mg arm (n = 37 patients)-adverse event [16], abnormal lab [2], unsatisfactory response [4], protocol violation [3], withdrew consent [4], administrative problem [1], graft loss [4], missing lab data [3]; and in the MPA arm (33 patients)-adverse event [11], unsatisfactory response [6], no longer requires study drug [1], protocol violation [1], withdrew consent [2], administrative problem [2], death [3], graft loss [5], missing lab data [2].

Table 4. Characteristics of study population at 3 months following transplant, stratified by proteinuria status and by treatment arm
 At month 3 post-TX:No proteinuria (<300 mg/g) N = 534Proteinuria (≥300 mg/g) N = 192EVR 1.5 No proteinuria N = 182EVR 1.5 Proteinuria N = 58EVR 3.0 No proteinuria N = 155EVR 3.0 Proteinuria N = 87MPA No proteinuria N = 197MPA Proteinuria N = 47
Recipient demographic summary
 Age (mean, SD)45.8, 12.7246.0, 14.0045.5, 12.245.1, 15.344.6, 13.345.1, 13.647.1, 12.748.9, 12.9
 Males (n, %)362 (67.8)123 (64.1)117 (64.3)36 (62.1)111 (71.6)55 (63.2)134 (68.0)32 (68.1)
 Caucasians (n, %)353 (65.9)143 (74.5)123 (67.6)44 (75.9)97 (62.6)62 (71.3)132 (67.0)37 (78.7)
 BMI (mean, SD)25.7, 4.926.0, 5.125.6, 4.925.7, 5.625.7, 5.125.7, 4.825.7, 4.726.9, 4.8
 0 HLA mismatches (n, %)21 (3.9)7 (3.6)5 (2.7)1 (1.7)6 (3.9)4 (4.6)10 (5.1)2 (4.3)
 Most recent PRA (mean, SD)1.3, 4.751.4, 5.081.8, 5.21.5, 5.51.4, 5.81.7, 5.50.8, 3.10.9, 3.5
Primary disease number (%)
 Glomerulonephritis/glomerular disease92 (17.2)35 (18.2)33 (18.1)8 (13.8)32 (20.6)18 (20.7)27 (13.7)9 (19.1)
 Pyelonephritis/interstitial nephritis16 (3.0)2 (1.0)6 (3.3)2 (3.4)3 (1.9)0 (0.0)7 (3.6)0 (0.0)
 Polycystic disease69 (12.9)20 (10.4)29 (15.9)2 (3.4)17 (11.1)8 (9.2)23 (11.7)10 (21.3)
 Hypertension/nephrosclerosis87 (16.3)37 (19.3)29 (15.9)13 (22.4)32 (20.6)13 (14.9)26 (13.2)11 (23.4)
 Diabetes mellitus61 (11.4)37 (19.3)22 (12.1)11 (19.0)9 (5.8)15 (17.2)30 (15.2)11 (23.4)
 Other186 (34.8)53 (27.6)52 (28.6)18 (31.0)58 (37.4)29 (33.3)76 (38.6)6 (12.8)
Donor characteristics summary
 Age (mean, SD)40.5, 13.4842.9, 12.7440.4, 13.941.8, 13.340.3, 13.342.2, 11.940.8, 13.345.4, 13.4
 Male (n, %)292 (54.7)78 (40.6)107 (58.8)25 (43.1)80 (51.6)39 (44.8)105 (53.3)14 (29.8)
 Caucasian (n, %)361 (67.6)143 (74.5)122 (67.0)43 (74.1)102 (65.8)62 (71.3)137 (69.5)38 (80.9)
Type n (%)
 Deceased donor224 (42.1)95 (50.0)79 (43.4)29 (50.0)60 (38.7)42 (48.2)86 (43.7)25 (53.2)
 Living related218 (40.8)65 (33.9)70 (38.5)19 (32.8)71 (45.8)33 (37.9)77 (39.1)65 (27.7)
 Living unrelated91 (17.0)31 (16.1)33 (18.1)10 (17.2)24 (15.5)12 (13.8)34 (17.3)9 (19.1)
Transplant information
 Cold ischemia time (hours) mean, SD7.6, 8.88.8, 9.17.4, 8.28.8, 9.17.1, 8.88.2, 8.38.2, 9.39.9, 10.3
 Delayed graft function (n, %)27 (5.1)20 (10.4)14 (7.7)5 (8.6)7 (4.5)9 (10.3)6 (3.0)6 (12.8)
 Acute rejection (n, %)34 (6.4)18 (9.4)13 (7.1)4 (6.9)7 (4.5)6 (6.9)14 (7.1)8 (17.1)
 Preemptive transplant (n, %)82 (15.4)33 (17.2)30 (16.5)8 (13.8)22 (14.2)13 (14.9)30 (15.2)12 (25.5)

The characteristics of those patients who had proteinuria versus those that did not are shown in Table 4, stratified by treatment arm. The three groups were similar in most characteristics, however, there was a higher percentage of preemptive transplant and a higher rates of acute rejection in the MPA cohort with proteinuria at 3 months. Pretransplant diabetes was associated with higher rates of proteinuria among all three groups.

When examining estimated GFR over the 24-month study period, those patients with early proteinuria consistently had lower eGFR at every subsequent time point (Table 5). This was more pronounced in the EVR 3.0 and MPA arms initially, but at 24 months the difference in GFR between those with versus without proteinuria (–9.6 mL/min overall; p < 0.001) were comparable between treatment arms.

Table 5. Relationship of early (3 months posttransplant) proteinuria to eGFR and graft loss, by treatment arm. p-Values are in comparison to the mycophenolate (MPA)-treated arms, by proteinuria status
Proteinuria status at month 3:EVR 1.5 No proteinuria N = 182EVR 1.5 proteinuria N = 58 (24%)EVR 3.0 No proteinuria N = 155EVR 3.0 proteinuria N = 87 (36%)MPA No proteinuria N = 197MPA proteinuria N = 47 (19%)
eGFR 3 months, mL/59.0 (19.5)57.1 (25.2)56.0 (17.0)49.3 (21.7)58.1 (18.9)41.8 (16.5)
 min (mean, SD)p = 0.660p < 0.001p = 0.285p = 0.054  
6 months59.1 (18.9)55.6 (21.1)57.2 (18.0)51.4 (19.1)55.9 (16.2)42.9 (16.7)
 p = 0.096p = 0.004p = 0.506p = 0.024  
12 months59.3 (16.9)57.1 (24.9)59.6 (18.7)50.6 (18.1)56.4 (16.3)47.4 (15.0)
 p = 0.115p = 0.047p = 0.114p = 0.400  
18 months59.8 (16.8)50.9 (22.2)60.8 (19.4)51.1 (17.5)56.4 (16.4)49.6 (17.6)
 p = 0.078p = 0.809p = 0.049p = 0.715  
24 months59.7 (19.6)52.2 (20.3)60.5 (18.2)49.2 (19.7)58.9 (20.5)49.3 (19.3)
 p = 0.775p = 0.589p = 0.530p = 0.970  
Graft survival at 2498%95%98%94%99%94%
 monthsp = 0.199p = 0.999p = 0.324p = 0.999  
Graft loss (n)433513
Proteinuria 300–1500 mg/g Cr 2 5 3
Proteinuria > 1500 mg/g Cr 1 0 0
Cause of graft lossCAN (2) AR (1) Sepsis (1)CAN (1) Bleeding postbiopsy (1) Prolif GN (1)CAN (2) Infection (1)CAN (2) IgAN (1) Nonadherence (1) PNF(1)Sepsis (1)CAN (2) Nonadherence (1)

Similar to the eGFR findings, 24-month graft survival was also lower in patients who exhibited proteinuria at month 3 (94%) than those who did not (99%; p = 0.003; Table 5). The causes of graft loss did not follow a specific pattern among treatment arms, with the diagnosis of chronic allograft nephropathy given in similar proportions among the three groups. A higher degree of proteinuria at 3 months was not significantly worse than mild/moderate proteinuria in its association with graft loss. When stratifying patients with proteinuria into 300–1500 mg/day versus >1500 mg/day cohorts, graft survival in those with proteinuria 300–1500 mg/g was 94.1% (p = 0.004 vs. those without proteinuria) whereas graft survival in patients with proteinuria >1500 mg/g was 95.5% (p = 0.307 vs. those without proteinuria, 98.5%), however, the sample size for those with higher degrees of proteinuria is small. Using logistic regression, the OR for graft loss was 3.8 (p = 0.005) for those with proteinuria versus no proteinuria at month 3 and again was not different between treatment groups (p = NS).

Finally, to address the natural progression of proteinuria over time by treatment arm, we analyzed changes in proteinuria over time stratified by proteinuria status at 3 months while also comparing study discontinuation rates and ACE/ARB use within groups (Table 6). For those with early proteinuria that remained in the analysis, median proteinuria and proteinuria rates ≥300 mg/g Cr at 1 and 2 years remained stable and were similar among the EVR arms, although falling in the MPA arm. In those without early proteinuria, the development of proteinuria 300 mg/g Cr at 2 years was noted in 19% and 22% of those in the EVR arms (with identical median proteinuria levels of 328 mg/g Cr) and was 7% in the MPA arm (with a median proteinuria of 200 mg/g Cr). Those who developed proteinuria after 3 months demonstrated lower eGFR than those without proteinuria at 24 months, although this difference was not as marked in the EVR 1.5 cohort. In those with early proteinuria, those with initial proteinuria but whose proteinuria regressed to < 300 mg/g Cr after 3 months had higher eGFR of 8 mL/min in the EVR cohorts, a finding that was not evident in the MPA arm. Taking these findings into account together with graft function data from Table 4, early proteinuria is associated with poorer eGFR and graft survival, and in those without early proteinuria the use of EVR is associated with two- to threefold increased rate of proteinuria over time compared to MPA. The development of proteinuria in the EVR 1.5 mg arm after 3 months was associated with a nominal (57 mL/min vs. 60 mL/min) difference in eGFR, with larger differences in eGFR in the MPA and EVR 3.0 arms.

Table 6. Evolution of proteinuria, eGFR- and study discontinuation over time in groups with and without proteinuria 300 mg/g at 3 months post-transplant. p-Values are in comparison to the mycophenolate (MPA)-treated arms, by proteinuria status
Proteinuria status  at month 3:EVR 1.5 No proteinuria N = 182EVR 1.5 proteinuria N = 58 (24%)EVR 3.0 No proteinuria N = 155EVR 3.0 proteinuria N = 87 (36%)MPA No proteinuria N = 197MPA proteinuria N = 47 (19%)
Median proteinuria at 1 year142 (91–235)323 (222–580)137 (86–233)287 (196–688)96 (69–158)299 (141–864)
 (mg/g Cr, interquartile range)p <0.001p = 0.965p <0.001p = 0.751  
Median proteinuria at142 (84–274)328 (134–966)134 (88–240)328 (194–653)107 (69–172)200 (96–382)
 2 years (mg/g Cr, interquartile range)p = 0.001p = 0.102p <0.001p = 0.034  
Patients with proteinuria27/148 (18%)21/37 (57%)23/120 (19%)26/55 (47%)17/168 (10%)14/30 (47%)
 ≥ 300 mg/g Cr at 1 year n (%)p = 0.050p = 0.467p = 0.038p = 0.999  
Patients with proteinuria28/132 (21%)14/27 (52%)21/109 (19%)22/42 (52%)10/140 (7%)8/26 (31%)
 ≥ 300 mg/g Cr at 2 years n (%)p <0.001p = 0.166p = 0.006p = 0.131  
eGFR at 2 years byYNYNYNYNYNYN
 proteinuria status at57 (23)60 (19)50 (17)58 (24)52 (16)63 (19)45 (20)53 (19)51 (18)60 (21)55 (24)49 (15)
 2 years, mL/min (mean, SD)0.4560.8830.6100.2090.8540.3240.2710.503    
Discontinued from study34 (19%)21 (36%)35 (23%)32 (37%)29 (15%)17 (36%)
 by 1 year n (%)p = 0.335p = 0.999p = 0.0702p = 0.999  
Discontinued from study50 (27%)31 (53%)46 (30%)45 (52%)57 (29%)21 (45%)
 by 2y n (%)p = 0.820p = 0.434p = 0.906p = 0.473  
Median proteinuria at last147 (91–288)608 (318–2251)196 (112–297)581 (287–1108)107 (71–224)517 (269–1519)
 measured endpoint for dropouts through 2 years (mg/g Cr, interquartile range)p = 0.053p = 0.473p = 0.002p = 0.967  
eGFR at last time point49 (18)39 (23)44 (16)47 (23)45 (14)29(13)
 before discontinuation, mL/min (mean, SD)p = 0.170p = 0.066p = 0.732p < 0.001  
ACEI use (at any time125 (69%)39 (67%)99 (64%)55 (63%)117 (59%)32 (68%)
 during study through 1 year) n (%)p = 0.069p = 0.999p = 0.441p = 0.705  
ACEI use (at any time132 (73%)41 (71%)107 (69%)55 (63%)122 (62%)33 (70%)
 during study through 2 years) n (%)p = 0.029p = 0.999p = 0.178p = 0.451  

Of the entire study population, 97/192 (51%) of patients with early proteinuria discontinued the study by 2 years posttransplant compared to 153/534 (29%) of those without early proteinuria (Table 6). Discontinuation rates were similar among treatment arms both for subjects with and without early proteinuria at 1 and 2 years, whereas median proteinuria levels before discontinuing the study were also similar among treatment arms, with the exception of the MPA arm without early proteinuria. ACEI/ARB use in the first year and at 2 years following transplant were similar not only among treatment arms but also among those with and without proteinuria at 3 months. These findings reduce the potential bias that may introduced by antiproteinuric medications and reduce the possibility that one treatment arm experienced a significantly higher dropout rate or had subjects with higher grades of proteinuria drop out of the analysis. However, proteinuric subjects in the MPA arm who discontinued follow-up had significantly lower eGFR than the other groups (29 mL/min vs. 39–49 mL/min in other groups), introducing the potential for bias of higher eGFR in the remaining proteinuric MPA cohort (i.e. patients with more severe renal dysfunction in the proteinuric MPA arm were excluded from “on therapy” analysis by the 24-month endpoint).


This post hoc analysis of a large multicenter randomized trial provides a number of novel observations related to the clinical significance of mild to moderate levels of proteinuria early following transplant, as well as the clinical implications of the association of mTOR inhibition and proteinuria. First, mTOR inhibitor use (in this study, EVR) is associated with a risk of proteinuria over standard immunosuppression that is dose dependent. Although higher trough concentrations of EVR > 8 ng/mL were associated with a 1.8-fold increased hazard for proteinuria ≥ 300 mg/day, EVR at 1.5 mg/day and adjusted to trough concentrations of 3–8 ng/mL imparted a nonsignificant 1.2-fold increased risk. Second, proteinuria ≥ 300 mg/g Cr identified at 3 months following transplant is associated with lower eGFR (–9.6 mL/min; p < 0.001) and graft loss (OR for graft loss 3.8; p = 0.005) at 24 months compared to those without early proteinuria. Interestingly, these findings were not more robust in the EVR treatment arms despite higher rates of proteinuria, leaving in question whether proteinuria from specific immunosuppressive agents or other etiologies of proteinuria may have contributed to graft outcomes over the 2-year follow-up period. Third, the development of proteinuria ≥ 300 mg/g Cr from 3 to 24 months posttransplant was two- to threefold higher with EVR compared to MPA, but its association with diminished eGFR was particularly noted in the EVR 3.0 and MPA cohorts. Finally, we find that over 70% of subjects on a CSA/EVR regimen did not develop proteinuria during this time, and approximately 25% underwent regression of proteinuria following identification at 3 months, again associated with higher eGFR than those that had persistent proteinuria. In total, these findings help to clarify the importance of utilizing EVR with low to moderate drug exposure, and the value of monitoring for proteinuria ≥ 300 mg/g Cr early as 3 months following transplant.

Varying degrees of proteinuria, measured at varying times posttransplant, have previously been shown to be associated with graft loss and death. Reviewed in [18], most studies have not examined the prevalence of proteinuria as low as > 300 mg/day, have not correlated proteinuria with GFR, or assessed proteinuria as early as 3 months posttransplant. No single study has performed all of these analyses, none have been performed using a multicenter cohort, and none have examined the impact of immunosuppression on these parameters. Smaller case-control studies and retrospective single center studies have demonstrated a similar relationship of proteinuria at 3 months with graft loss [19][21], and a large single center study correlated degrees of proteinuria as low as 150 mg/day at 12 months with findings on protocol biopsy and with graft loss [3]. In total, our analysis provides the best evidence to date that proteinuria > 300 mg/day as early as 3 months posttransplant may be an excellent risk stratification tool for graft outcomes and provide a framework for future interventional studies.

From a clinical perspective, these data help to understand the actual as well as the as-yet poorly defined risks of mTOR therapy with respect to the development of proteinuria and its potential contribution to graft dysfunction. In this study, proteinuria is associated with diminished GFR, regardless of etiology or immunouppressive regimen. We report the incidence of proteinuria with two FDA-approved treatment regimens (CSA/MPA and CSA/EVR) and note a two- to threefold increased risk of the incidence of de novo proteinuria by 2 years following transplant with CSA/EVR. However, it is of interest that over 70% of subjects on CSA/EVR did not develop persistent proteinuria during this time, that a significant proportion of those with early proteinuria and follow-up to 2 years underwent regression of proteinuria, that both in early and developing proteinuria a consistent association of proteinuria and lower GFR was noted, with comparable eGFR to CSA/MPA-treated subjects. Given these data, if there are clinical considerations to utilize a CSA/EVR regimen, a reasonable approach may be to adjust EVR exposure to trough levels 3–8 ng/mL, measure proteinuria at 3 months and periodically, with conservative antiproteinuric strategies (e.g. ACEI/ARB use) employed when proteinuria is identified and consider discontinuation of mTOR if proteinuria is not responsive to therapy or is progressive.

An important distinction to be made with this study when comparing rates of proteinuria with mTOR use is that EVR was used in combination with CNI, unlike a number of prior descriptions of proteinuria associated with mTOR in the absence of CNI. In studies with 2-year follow-up, the Spare the Nephron trial participants on SRL (goal 5–10 ng/mL; [11] had a mean urinary protein/creatinine ratio of 0.6 at 24 months (estimated proteinuria of 600 mg/day) versus 0.39 in this study. In the Zeus trial, participants on EVR (goal 6–10 ng/mL; [12] had 17% of its participants with proteinuria > 500 mg/day at 24 months (similar to the 17% rate in this study). In the CONVERT trial, participants on SRL (goal trough 8–20 ng/mL; [14] had median urinary protein/creatinine ratio of 0.33 at 24 months (vs. 0.15 in this study). Although study participants differ in baseline risks and mTOR exposure among these trials, these data lend support to the concept that the risk of proteinuria with mTOR may be less when used in the de novo setting in combination with CNI than following transition to a CNI-free mTOR-based regimen.

This analysis has a number of limitations to consider, beyond the biases that may be inherent to the study design. Study design limitations include the “on therapy” nature of assessment that is required for the time-dependent models (i.e. patients must be receiving a medication with a drug level and a proteinuria assessment available to be included in this component of the analysis). Although ACEI/ARB use was consistent among groups over the entire study, residual confounding from the possible use of ACEI/ARB may exist because of the inability to associate timing of ACEI/ARB use in the time-dependent model. For the assessment of outcomes of patients stratified by proteinuria status at 3 months, limitations include (i) the lack of biopsy data that may identify other causes of proteinuria such as recurrent primary renal disease, donor disease, or the impact of prior episodes of acute rejection or delayed graft function; and (ii) the assumption that proteinuria at 3 months was specifically from the transplant (and not native) kidney. There was a higher rate of preemptive transplant in the CSA/MPA arm that had early proteinuria and we cannot exclude the possibility that this was a result of residual native proteinuria, despite studies that have demonstrated a regression by 3 months posttransplant. There were no baseline proteinuria assessments before 3 months to assist in distinguishing this further. Finally, this study includes data only to 24 months and cannot predict the impact of the higher levels of proteinuria noted in EVR arms (Figure 1) upon longer term outcomes.


In conclusion, this study represents the largest and most comprehensive study to correlate a dose-dependent effect of EVR with the risk of proteinuria. Together with the previously reported clinical efficacy of EVR with trough level goals of 3–8 ng/mL and low-dose CSA, these data clarify a therapeutic target range of EVR that reduces but does not eliminate proteinuria risk. Regardless of immunosuppressive regimen, this study identified proteinuria ≥ 300 mg/g at 3 months posttransplant to be associated with lower eGFR and graft survival at 2 years, a potentially important marker for intervention and future study.


The authors wish to thank all the 2309 investigators who contributed to this study.

Funding: for this study was provided by Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA.


The authors of this manuscript have conflicts of interest to disclose as described by the American Journal of Transplantation. ACW is a Consultant for Novartis, Tolera, and Bristol Meyers Squibb, and has received Speaker honoraria for Novartis. FG was an employee of Novartis Pharmaceuticals Corporation at the time the study was conducted. KM is an employee of Novartis Pharmaceuticals Corporation. YK was an employee of Novartis Pharmaceuticals Corporation at the time the study was conducted. MC is a Consultant and has received speaker honoraria for Novartis.