Tacrolimus Exposure and Evolution of Renal Allograft Histology in the First Year After Transplantation

Authors


* Corresponding author: Dirk R.J. Kuypers, Dirk.Kuypers@uz.kuleuven.ac.be

Abstract

Tacrolimus has a narrow therapeutic window and is characterized by a large inter-individual variability in bioavailability. The impact of tacrolimus exposure on subclinical evolution of graft histology has not been studied in renal recipients. This analysis included 239 protocol biopsies (obtained at implantation, 3 and 12 months) of 120 consecutive kidney recipients treated with tacrolimus, mycophenolate mofetil (MMF) and corticosteroids. Biopsies were scored according to the Banff 2001 criteria and a chronicity score was calculated. Prospective pharmacokinetic data were included in the analysis (5544 tacrolimus predose blood concentrations and tacrolimus AUC0-12 at 3 and 12 months). Higher donor age and higher number of human leukocyte antigen-DR (HLA-DR) mismatches were independent predictors of subclinical acute rejection at 3 months, present in 8.7% of patients. The number of HLA-DR mismatches was independently associated with biopsy-proven clinical acute rejection. Biopsy-proven acute rejection episodes and low mean tacrolimus exposure were independently associated with higher increase in chronicity scores between 3 and 12 months after transplantation. This observational study suggests that rejection phenomena and immune-mediated mechanisms remain important in the early progression of chronic allograft pathology. Tacrolimus doses or systemic exposure were not associated with lesions of calcineurin inhibitor nephrotoxicity, suggesting that other factors determine susceptibility to tacrolimus nephrotoxicity.

Introduction

Despite a clear improvement in short-term graft survival over the past decades, long-term graft survival has increased to a lesser extent (1). After death with a functioning graft, chronic allograft nephropathy is the major cause of graft loss characterized by interstitial fibrosis and tubular atrophy (IF/TA) (2). Tubular atrophy and interstitial fibrosis are non-specific lesions, and represent a final common pathway of different types of damage to the transplanted kidney, including calcineurin inhibitor nephrotoxicity (CNIT) and immunologic injury (3,4). Longitudinal analysis using protocol biopsies have shown that chronic lesions often originate already very early after transplantation with subclinical acute rejection (SAR) episodes contributing to this chronic damage (5,6).

Although both calcineurin inhibitors cyclosporine A (CsA) and tacrolimus are characterized by a narrow therapeutic window and a large inter-individual variability in bioavailability, few studies have addressed the optimal long-term target levels in order to avoid CNIT (7). One study has suggested that higher CsA doses and higher CsA predose trough levels are associated with more severe arteriolar hyalinosis lesions, suggestive of CNIT (8). In contrast, lower CsA C0 and C2 levels have been associated with progression of IF/TA and more severe chronic histological lesions (9,10). In tacrolimus-treated patients no such data are available.

The current prospective study was undertaken in patients treated with tacrolimus, mycophenolate mofetil (MMF) and corticosteroids in order to shed more light on the different clinical factors associated with the development of chronic allograft pathology in the first year after transplantation and focusing on the impact of tacrolimus exposure on this evolution.

Methods

Subjects

One hundred twenty patients who received a single kidney transplant at our institution between the 1st of March 2004 and the 7th of May 2006, and treated with an immunosuppressive regimen of tacrolimus (Prograft® Fujisawa GmbH) in combination with MMF (CellCept® Roche Diagnostics, Mannheim, Germany) and oral methylprednisolone (Medrol® Upjohn, Puurs, Belgium) consented to participate in this study. Combined transplant recipients were excluded. Thirty patients with higher immunologic risk received additional induction therapy with basiliximab (n = 23), daclizumab (n = 1) or anti-thymocyte globulines (n = 6).

Tacrolimus and MMF were administered twice daily, starting on the day of transplantation. The daily tacrolimus dose was adjusted to achieve target 12-h predose blood concentrations between 12 and 15 ng/mL in the first 3 months after transplantation. Thereafter, doses were adjusted to achieve predose concentrations of 9–12 ng/mL. Doses of MMF were reduced only on clinical indication (side effects).

At 3 and 12 months after transplantation, patients underwent a protocol biopsy. At 3 months, 118 of the consenting patients underwent a protocol biopsy, three patients were excluded from the analysis because of insufficient data on tacrolimus exposure, leaving 115 patients for this analysis. At 12 months, a protocol biopsy was performed in 63 of the 120 subjects. The number of biopsies performed at 12 months was lower than at 3 months because follow-up of most recent transplanted patients was shorter than 1 year. Renal biopsy cores were obtained by an automated biopsy gun using an 18-gauge needle and real-time ultrasound guidance. A baseline biopsy obtained at implantation was available in 61 cases.

Histological evaluation

Slides containing 4–10 paraffin sections (2 μm) were stained with hematoxylin eosin and with a silver methenamine staining method (Jones). An immunohistochemical C4d stain (monoclonal antibody, dilution 1:500, Quidel Corporation, Santa Clara, CA) was performed on frozen tissue. Histological lesions were semi-quantitatively scored according to the revised Banff 2001 criteria (2,11,12). All biopsies (n = 239; 61 at implantation, 115 at 3 months and 63 at 1 year) were reviewed blindly (EL). To assess changes in histological lesions, ‘delta scores’ were calculated based on the semi-quantitative Banff scores (0–3) at 3 and 12 months posttransplantation. A chronicity score was calculated as the sum of four basic ‘chronic’ Banff qualifiers (chronic glomerular damage [cg], interstitial fibrosis [ci], tubular atrophy [ct], vascular intimal thickening [cv], thus allowing for a total score ranging from zero to a maximum score of 12 (13).

All patients with clinical and subclinical Banff type I or II acute rejection were treated with high doses of methylprednisolone in a tapering protocol. Subclinical borderline rejection was not treated. The result of the 3-month protocol biopsy was not used to subsequently adjust future tacrolimus target levels or MMF dose, irrespective of the histological findings, except in case of BK-polyomavirus associated interstitial nephritis (n = 2).

Pharmacokinetics

For the pharmacokinetic assessments at 3 and 12 months (on the day of the protocol biopsy), patients had to adhere to an overnight fast for at least 10 h and the morning dose of MMF and tacrolimus were ingested 12 h after the previous dose. Patients had to be on a stable immunosuppressive regimen at the time of pharmacokinetic assessment (i.e. dose unchanged in the week previous to the pharmacokinetic analysis). Immediately before tacrolimus and MMF dosing (C0) and after 30, 60, 90 min and 2, 3 and 4 h blood samples were taken. Furthermore, all tacrolimus predose trough levels determined in the first year after transplantation were included in the analysis, as well as tacrolimus doses. In total, 4290 predose tacrolimus levels between 0 and 3 months were included (mean 36.4 ± 7.8 per patient) and 1254 measurements (mean 19.9 ± 12.3) from 3 to 12 months after transplantation. Determination of blood concentrations of tacrolimus was performed by using the Tacrolimus II MEIA/IMx assay (Abbott Laboratories). Dose-interval tacrolimus exposure at 3 and 12 months was calculated using a previously validated algorithm using five sampling points. This algorithm explained 96% of the variance in dose-interval tacrolimus AUC0-12 (14).

Data collection and statistical analysis

All data were collected prospectively. The mean serum creatinine levels were calculated per patient in the time intervals 0–3 and 3–12 months and also mean doses of methylprednisolon and mean arterial blood pressure were calculated from the electronic patient records in these time intervals. Data analysis was performed using SAS software (SAS 8.2 and Enterprise Guide 1.3; SAS institute). One-way ANOVA or Kruskal-Wallis test and linear regression were performed for unpaired continuous variables, and the Chi-square test or Fisher's exact test to assess the association between categorical data, as appropriate. Correlations between continuous variables were assessed by Pearson correlation, between ordinal variables Spearman correlations were calculated. For comparison of paired data (the intra-individual histological evolution between time points), the paired t-test was used. As 61 paired zero and 3-month biopsies were available, the paired comparison between these biopsies was performed on 61 patients. The clinical determinants of 3 months histology were assessed using all 115 patients. Likewise, 63 patients (all patients with paired biopsies at 3 and 12 months) were considered in the analysis for determinants of the histology at 12 months. Multiple logistic regression analysis with backward selection was used to model the risk factors for the different histological patterns. Parameters with a p value < 0.2 in univariate analysis were included in the multivariate model. Two-sided p-values less than 0.05 were considered significant. Data are expressed as mean and standard deviation, unless stated otherwise.

Results

Patient and donor demographics and transplantation-related characteristics are summarized in Table 1.

Table 1.  Patient and donor demographics, transplantation characteristics of patients included in the 3 months analysis (n = 115)
ParameterMean ± SD
  1. 1Defined as the need for dialysis in the first week after transplantation.

  2. 2Diagnosed on an indication biopy performed in response to a rise in serum creatinine concentration.

Recipient age (years)53.8 ± 13.5
Recipient gender (% male)60.90%
Donor age (years)44.9 ± 16.4
Donor gender (% male)58.30%
Stroke as reason of donor death46.10%
Retransplantantion (%)19.10%
Diabetes prior to transplantation (%) 7.80%
Induction therapy (%)26.10%
Cold ischemia time (h)15.3 ± 8.41
Delayed graft function1 (%) 9.60%
Total number of HLA mismatches (N)2.61 ± 1.38
 Number of HLA-A mismatches (N)0.84 ± 0.71
 Number of HLA-B mismatches (N)1.02 ± 0.65
 Number of HLA-DR mismatches (N)0.77 ± 0.57
Biopsy-proven acute rejection in the first year2 (%)20.90%
Tacrolimus C0 0–3 months (ng/mL)13.4 ± 1.00
Tacrolimus C0 3–12 months (ng/mL)10.7 ± 1.70

Evolution of histological parameters in the first year after transplantation (Table 2)

Table 2.  The evolution of subclinical histology in the first year after transplantation between implantation and 3 months (A) and between 3 and 12 months (B)
 Baseline n = 613 Months n = 61p-Value1 0 vs. 3 months
  1. 1Paired t-testing between baseline and 3 months for 61 patients.

  2. 2Chi-square test for association.

  3. 3Paired t-testing between 3 and 12 months for 63 patients.

Interstitial fibrosis (‘ci’ score) 0.25 ± 0.570.48 ± 0.790.0219
Tubular atrophy (‘ct’ score) 0.49 ± 0.570.67 ± 0.570.0473
Vascular intimal thickening (‘cv’ score)0.082 ± 0.330.57 ± 0.870.0001
Transplant glomerulopathy (‘cg’ score)0.016 ± 0.130.0 ± 0.0NS
Arteriolar hyalinosis (‘ah’ score) 0.23 ± 0.500.49 ± 0.790.0224
Mesangial matrix increase (‘mm’ score)0.066 ± 0.250.26 ± 0.660.0221
% sclerosed glomeruli0.055 ± 0.120.048 ± 0.12NS
Subclinical acute rejection (SAR) (%)9.26%
Subclinical borderline rejection (%)9.26%
Chronic allograft nephropathy (IF/TA) (%)17.80%20.37%NS2
 Banff grade 117.80%16.67%NS2
 Banff grade 2–30.00%3.70%NS2
Chronicity score0.84 ± 1.081.72 ± 1.68<0.0001
 3 Months n = 6312 Months n = 63p-Value3 3 vs. 12 Months
Interstitial fibrosis (‘ci’ score)0.38 ± 0.630.76 ± 0.910.0012
Tubular atrophy (‘ct’ score)0.62 ± 0.521.16 ± 0.63< 0.0001
Vascular intimal thickening (‘cv’ score)0.59 ± 0.840.84 ± 1.04NS
Transplant glomerulopathy (‘cg’ score)0.0 ± 0.00.095 ± 0.43 NS
Arteriolar hyalinosis (‘ah’ score)0.56 ± 0.780.63 ± 0.79NS
Mesangial matrix increase (‘mm’ score)0.40 ± 0.790.17 ± 0.52NS
% sclerosed glomeruli0.024 ± 0.0560.030 ± 0.051NS
Subclinical acute rejection (SAR) (%)9.52%1.60%NS2
Subclinical borderline rejection (%)11.11%6.40%NS2
Chronic allograft nephropathy (IF/TA) (%)23.81%52.40%0.00092
 Banff grade 123.81%34.90%0.04062
 Banff grade 2–30.00%17.50%0.00012
Chronicity score1.59 ± 1.442.89 ± 2.28< 0.0001

Chronic histological lesions increased in the first 3 months after transplantation and subsequently as well from 3 months to 1 year, a significant increase in IF/TA scores was observed. Increase in vascular intimal thickening (‘cv’ score) was most prominent in the first 3 months, as was the increase in arteriolar hyalinosis (‘ah’) lesions and mesangial matrix (‘mm’). Consequently, the chronicity score increased significantly between implantation and 3 months, and between 3 and 12 months after transplantation. Twenty-one percent (20.9%) of patients developed biopsy-proven clinical acute rejection (BPAR), 92% of these within the first 2 weeks after transplantation. All biopsy-proven acute rejection episodes were successfully treated with high doses of methylprednisolone.

Subclinical acute rejection at 3 months

The prevalence of subclinical cellular infiltrates was highest at 3 months after transplantation (17.4%): SAR was diagnosed in 8.7% of biopsies while 8.7% of biopsies at 3 months were categorized as borderline rejection. It is of note that eight out of 10 episodes of SAR at 3 months had a vascular component. In addition, SAR was associated with C4d positivity: 50% of 3-months biopsies with SAR (5/10) were positive for C4d in the peritubular capillaries in contrast to only 5.7% (6/105) of biopsies without SAR (p < 0.0001). Clinical determinants associated with SAR are shown in Table 3. In multivariate logistic regression analysis, donor age and the number of HLA-DR mismatches were significant and independent predictors of SAR at 3 months after transplantation (Table 3, Figure 1). Patients with SAR at 3 months had significantly higher Banff scores for IF/TA already at implantation (p = 0.001 and p = 0.03, respectively). Consequently, the chronicity score of the baseline biopsy was significantly higher in patients experiencing SAR at 3 months (p = 0.01). Also BPAR was associated with higher number of HLA-DR mismatches (p = 0.0005) and cold ischemia time (p = 0.0043) in univariate analysis. Donors of patients with BPAR tended to be older (50.4 ± 14.7 vs. 43.4 ± 16.6; p = 0.06) and to have higher chronicity scores of the implantation biopsy (1.22 ± 1.31 vs. 0.67 ± 0.94; p = 0.08). In multivariate analysis, only the number of HLA-DR mismatches remained in the final model as independent predictor of BPAR (Odds ratio per mismatch = 3.73; 95% CI 1.22–11.5). Creatinine levels or creatinine clearance were not different between patients with and without SAR. There was no association between SAR and pretransplant panel-reactive antibody (PRA) levels or posttransplant presence of HLA antibodies. There were neither differences in the use of additional induction therapy or in exposure to MPA or tacrolimus in the 3 months preceding the protocol biopsy, nor in the last month immediately prior to this biopsy, comparing patients with versus without SAR.

Table 3.  Clinical parameters associated with subclinical acute rejection (SAR) at 3 months after transplantation in univariate and multivariate logistic regression analysis
Clinical ParametersNo SAR at 3 Months n = 105SAR at 3 Months n = 10Univariate p-ValueMultivariate p-Value
Recipient age (years)53.1 ± 13.561.1 ± 11.10.0704NS
Recipient gender (% male)61%60%NS 
Donor age (years)43.3 ± 16.062.0 ± 9.350.00050.0263 (OR 1.1; 1.0–1.15)
Donor gender (% male)59%50%NS 
Gender mismatch (female -> male)23.80%40%NS 
Donor stroke (%)44.80%60%NS 
Retransplantation (%)19%20%NS 
Number of HLA mismatches (N)2.50 ± 1.323.70 ± 1.570.0223NS
 HLA-A mismatches (N)0.84 ± 0.720.90 ± 0.57NS 
 HLA-B mismatches (N)0.99 ± 0.631.30 ± 0.820.145NS
 HLA-DR mismatches (N)0.70 ± 0.521.50 ± 0.53<0.0010.023 (OR 13.6; 2.5–76)
Cold ischemia time (h)15.5 ± 8.6013.5 ± 6.15NS 
Induction therapy (%)26.70%20%NS 
Delayed graft function (%)7.60%30%0.0215NS
Tacrolimus C0 0–3 months (ng/mL)13.3 ± 0.9413.9 ± 1.36NS 
Biopsy-proven acute rejection episode20%40%0.1429NS
Figure 1.

Independent association of the number of HLA-DR mismatches with subclinical acute rejection (SAR) at 3 months after transplantation.

Chronicity score

During the first year after transplantation, there was a steady increase in the calculated chronicity score (Table 2). Both at 3 and 12 months after transplantation, the calculated chronicity score correlated with creatinine clearance calculated with the formula of Cockcroft and Gault (respectively p = 0.003 and p < 0.0001, Figure 2) (15). Also the increase in chronicity score between 3 and 12 months correlated significantly with the creatinine clearance at 12 months (p = 0.003). The chronicity score could not be calculated in one biopsy at 3 months due to the absence of vascular structures, leaving 60 patients for paired analysis between implantation and 3 months.

Figure 2.

Correlation between chronicity score and simultaneous creatinine clearance both at 3 (A) and 12 (B) months.

In multivariate logistic regression analysis, both higher donor age (p = 0.01; OR per year increase 1.07; 1.01–1.12) and delayed graft function (p = 0.04; OR 5.836; 1.077–31.624) were independently associated with the increase in chronicity score between implantation and 3 months (delta chronicity score ≤ 1; n = 43 vs. ≥ 2; n = 17). As a result, the same variables (delayed graft function and donor age) were independently associated with the chronicity score at 3 months (Table 4).

Table 4.  Clinical parameters associated with high- vs. low-chronicity score at 3 months after transplantation in univariate and multivariate logistic regression analysis
Clinical ParametersChronicity Score < 3 at 3 Months n = 88Chronicity Score ≥ 3 at 3 Months n = 27Univariate p-ValueMultivariate p-Value
Recipient age (years)52.9 ± 14.356.6 ± 10.2NS 
Recipient gender (% male)60.2%63.0%NS 
Donor age (years)41.6 ± 16.455.8 ± 10.7<0.00010.0009 (OR 1.06; 1.03–1.10)
Donor gender (% male)55.7%66.7%NS 
Donor stroke (%)40.9%63.0%0.0502NS
Retransplantation (%)19.3%18.5%NS 
Number of HLA mismatches (N)2.53 ± 1.452.85 ± 1.13NS 
 HLA-A mismatches (N)0.81 ± 0.690.96 ± 0.76NS 
 HLA-B mismatches (N)1.00 ± 0.661.07 ± 0.62NS 
 HLA-DR mismatches (N)0.75 ± 0.570.81 ± 0.56NS 
Cold ischemia time (h)15.2 ± 9.2 15.8 ± 5.3 0.1903NS
Induction therapy (%)27.3%22.2%NS 
Delayed graft function (%)4.5%25.9%0.00330.0221 (OR 5.130; 1.27–20.8)
Tacrolimus C0 0–3 months (ng/mL)13.3 ± 1.0213.5 ± 0.90NS 
Biopsy-proven acute rejection episode (%)20.5%25.9%NS 

The clinical factors that were associated with the increase in chronicity score between implantation and 3 months were no longer associated with the increase in chronicity score between 3 and 12 months. In multivariate analysis, higher donor age and low exposure to tacrolimus were independently associated with high chronicity in the 12-month biopsies (Table 5). Assessing the increase in chronicity score between 3 and 12 months in multivariate ordered logistic regression, biopsy-proven acute rejection (BPAR) and low exposure to tacrolimus in this time interval were independently associated with higher increase in chronicity. Subclinical interstitial inflammation and tubulitis in the 3-months protocol biopsies were associated with higher increase in chronicity, but this was only a statistical trend in univariate analysis. In multivariate analysis, subclinical inflammation at 3 months was no longer present in the final model explaining variability in delta chronicity between 3 and 12 months (Table 6).

Table 5.  Clinical parameters associated with high- vs. low-chronicity score at 12 months after transplantation in univariate and multivariate logistic regression analysis
Clinical ParametersChronicity Score < 3 at 12 Months n = 34Chronicity Score ≥ 3 at 12 Months n = 28Univariate p-ValueMultivariate p-Value
Recipient age (years)52.2 ± 12.157.1 ± 14.00.0736NS
Recipient gender (% male)64.7%67.9%NS 
Donor age (years)39.4 ± 15.853.0 ± 12.40.00090.0033 (OR 1.07; 1.02–1.12)
Donor gender (% male)67.6%50.0%0.1979NS
Donor stroke (%)41.2%57.1%NS 
Retransplantation (%)20.6%17.9%NS 
Number of HLA mismatches (N)2.59 ± 1.422.79 ± 1.34NS 
 HLA-A mismatches (N)0.91 ± 0.710.86 ± 0.65NS 
 HLA-B mismatches (N)1.00 ± 0.651.07 ± 0.66NS 
 HLA-DR mismatches (N)0.68 ± 0.590.86 ± 0.65NS 
Cold ischemia time (h)14.9 ± 4.714.8 ± 6.7NS 
Induction therapy (%)26.5%21.4%NS 
Delayed graft function (%)5.9%17.9%NS 
Biopsy-proven acute rejection episode (%)11.8%35.7%0.034NS
Tubulitis at 3 months (‘t’ score)0.29 ± 0.580.46 ± 0.79NS 
Interstitial inflammation at 3 months (‘i’ score)0.24 ± 0.500.36 ± 0.73NS 
Vasculitis at 3 months (‘v’ score)0.029 ± 0.170.11 ± 0.31NS 
Tacrolimus C0 0–3 months (ng/mL)13.7 ± 0.8313.2 ± 0.850.0192NS
Tacrolimus C0 3–12 months (ng/mL)11.3 ± 1.439.95 ± 1.760.00450.0107 (OR 1.838; 1.152–2.933)
Table 6.  Clinical parameters associated with delta chronicity score between 3 and 12 months after transplantation in univariate and multivariate ordered logistic regression analysis
Clinical ParametersRelationUnivariate p-ValueMultivariate p-Value
Recipient age (years) NS 
Recipient gender (% male) NS 
Donor age (years)pos0.1734NS
Donor gender (% male)F>M0.0119NS
Donor stroke (%) NS 
Retransplantation (%) NS 
Number of HLA mismatches (N) NS 
 HLA-A mismatches (N) NS 
 HLA-B mismatches (N) NS 
 HLA-DR mismatches (N) NS 
Cold ischemia time (h) NS 
Induction therapy (%) NS 
Delayed graft function (%) NS 
Biopsy-proven acute rejection episode (%)pos0.02410.0347 (OR 3.199; 1.087–9.412)
Tubulitis at 3 months (‘t’ score)pos0.056NS
Interstitial inflammation at 3 months (‘i’ score)pos0.0562NS
Vasculitis at 3 months (‘v’ score)pos0.1525NS
Tacrolimus C0 0–3 months (ng/mL)neg0.0372NS
Tacrolimus C0 3–12 months (ng/mL)neg0.01460.0148 (OR 1.398; 1.068–1.830)

In multivariate analysis, low-tacrolimus exposure between 3 and 12 months was independently associated with higher increase in chronicity score in this time frame and subsequent higher chronicity scores at 12 months (see above). When comparing patients with relatively low mean exposure to tacrolimus (<9 ng/mL; n = 9) with patients having high exposure (>12 ng/mL; n = 14) between 3 and 12 months, the low group—despite having a similar chronicity profile at 3 months compared to the high group (chronicity score 1.78 ± 1.64 vs. 1.57 ± 1.34; p = NS)—had a significantly higher increase in chronicity score in this time interval and a higher chronicity score at 12 months than the high tacrolimus group (respectively p = 0.04 and p = 0.009; Figure 3). The correlation between tacrolimus exposure and delta chronicity score was −0.30 (p = 0.016). This association between low- tacrolimus exposure and high chronicity was mainly due to significantly higher scores for vascular intimal thickening and glomerulopathy in the low-exposure group (p = 0.03 and p = 0.02, respectively), but also the scores for interstitial fibrosis tended to be higher in this group (p = 0.10). There was no association between the measured tacrolimus AUC0-12 at 3 or 12 months and the histological parameters in the protocol biopsies at these time points or the histological evolution between these biopsies.

Figure 3.

Comparison of the chronicity score at 3 and 12 months, and of the increase in chronicity score in this time interval between low (<9 ng/mL), standard (9–12 ng/mL) and high (>12 ng/mL) mean tacrolimus C0 between 3 and 12 months after transplantation. Data are expressed as mean ± SEM. *p < 0.05; **p < 0.01 for low- vs. high-tacrolimus exposure groups.

The use of induction therapy in itself was not associated with differences in the histological appearance or evolution of the graft. However, when analyzing the clinical determinants of delta chronicity score between 3 and 12 months in the subgroup of patients who did not receive induction therapy (n = 47), a similar statistically significant and independent association of low-tacrolimus exposure with higher increase in chronicity was found. The negative correlation between mean tacrolimus exposure and increase in chronicity score was stronger in this subgroup (r =−0.45; p = 0.001). However, this was not the case in the patients who received additional induction therapy (n = 15). In this higher risk group no association was found between mean tacrolimus predose levels and increase in chronicity score.

Increase in arteriolar hyalinosis

There was no association between donor age, donor gender or donor body mass index and the presence (n = 12/61) of arteriolar hyalinosis at baseline. Between implantation and 3 months after transplantation, there was a significant increase in the overall Banff score for arteriolar hyalinosis (Table 2A). In multivariate analysis, only stroke as cause of donor death was independently associated with new onset or increase in arteriolar hyalinosis from implantation to 3 months after transplantation (p = 0.04; OR 3.47; 1.04–11.5). Presence of preexisting diabetes mellitus, mean blood pressure or the use of angio-tensin-converting enzymes inhibitors were not associated with delta ‘ah’ scores. None of the clinical determinants, including tacrolimus exposure parameters, was associated with new onset arteriolar hyalinosis at 12 months or with the increase in arteriolar hyalinosis score from 3 to 12 months. Likewise, in the absence of arteriolar hyalinosis lesions at 3 months after transplantation, exposure parameters of tacrolimus were not different between patients who developed new-onset arteriolar hyalinosis (n = 14) versus patients who did not develop arteriolar hyalinosis by 12 months (n = 24), nor were there differences in arterial blood pressure in this time period. Patients with new-onset arteriolar hyalinosis had not more preexisting or posttransplant diabetes neither were absolute HbA1c levels different between these groups.

The variability in mean tacrolimus exposure is not the reflection of therapeutic interventions. There was no association between mean tacrolimus exposure between 3 and 12 months and the increase in arteriolar hyalinosis scores in the first 3 months after transplantation or with the presence of arteriolar hyalinosis at 3 months, nor was there any association between mean tacrolimus exposure between 3 and 12 months and other clinical determinants in the first 3 months after transplantation (biopsy-proven acute clinical rejection episodes, graft function, donor and recipient characteristics, HLA matching) or with the histological findings at implantation and 3 months.

Discussion

This longitudinal biopsy study demonstrates that in patients treated with the combination of tacrolimus, MMF and corticosteroids, an important increase in chronic histological lesions develops already in the first year after transplantation. This study for the first time correlates the evolution of subclinical allograft histology to tacrolimus exposure, in a first attempt to search for optimal long-term tacrolimus target levels, and to identify the ‘typical’ histologic appearance of renal tissue exposed to different tacrolimus levels.

This study confirms that at least two distinct phases in the evolution of chronic histological damage exist (5): in the earliest time period (from implantation to 3 months after transplantation), donor age and delayed graft function are associated with a less favorable subclinical evolution, which is already reflected by poor allograft function at 3 months. From 3 to 12 months, additional histological damage occurs independent of donor-related factors. Chronic histological injury in the first 3 months after transplantation can be considered a consequence of cumulative effects of tissue damage from ischemia-reperfusion (I/R) (16,17) and possibly CNIT due to higher tacrolimus exposure in this time period. However, when assessing the true impact of tacrolimus exposure on complete graft histology, tacrolimus exposure alone in this early time after transplantation is not predictive of the histological evolution (in contrast to the period 3–12 months, see below), which suggests an overwhelming impact of I/R injury or other immediate peri-transplant factors.

The higher susceptibility of older kidneys to I/R injury is not the only explanation for the observed increases in chronic pathology with donor age (18), but also the association of higher donor age with inflammation in the graft (both BPAR as subclinical inflammation) plays a role. We indeed demonstrated that higher donor age and baseline histology is an independent risk factor for subclinical inflammation at 3 months, which is associated with higher increases in chronic graft damage later (19–21). It cannot be excluded that the subclinical infiltrates detected at 3 months are already present much earlier after transplantation, thus contributing to the early increase in chronic pathology, starting from implantation. In addition, this study confirms the association of both BPAR and SAR with a higher HLA-DR mismatch (22), suggesting that not only non-specific inflammatory events, but also donor-specific immune responses are playing a role in these inflammatory processes and that matching for HLA loci is still important, even in this era of powerful immunosuppressive regimens. Tacrolimus exposure was not associated with the presence of inflammation in the graft, neither BPAR nor subclinical inflammation.

In the second phase of this study, between 3 and 12 months, factors associated with the increase in chronicity scores are totally different. Donor-related parameters were no longer predictive for this increase—in contrast to the first 3 months after transplantation—but earlier inflammatory phenomena (BPAR and SAR, see above) played a significant and independent role in this later time frame. In addition, as CNIT is considered a major contributor to chronic transplant pathology, it could be conceived that higher calcineurin inhibitor drug levels are independent predictors of this increase in chronic allograft pathology. However, the current study demonstrates that higher doses or levels of tacrolimus were not associated with higher increases in lesions suggestive of CNIT. On the contrary, this study demonstrates for the first time that lower exposure to tacrolimus between 3 and 12 months after transplantation is associated independently with higher increases in chronic pathology.

As there was no association between mean tacrolimus exposure between 3 and 12 months and clinical or histological determinants in the preceding time interval or the protocol biopsy results at implantation and 3 months, the variability in mean tacrolimus exposure is not the reflection of therapeutic interventions, but the result of the high inter- and intra-individual variability in tacrolimus pharmacokinetics. Likewise, it should not surprise that mean tacrolimus C0 levels in any given time frame were relatively better predictors because they not only correlated well with tacrolimus AUC0-12 (R2= 0.71) (14), but were more frequently measured while tacrolimus AUC0-12 were only measured at two time points in this study, and thus subject to relatively large intra-individual variability.

As this study could not demonstrate a significant association between the use of induction therapy per se and the subclinical histological evolution, it could be suggested that this illustrates that induction therapy—which was used in high-risk recipients—does help to overcome this high risk and is able to maintain a similar histological evolution as in low-risk patients. Moreover, in a subanalysis, the association between low-tacrolimus exposure and higher increase in chronicity could only be found in the group not receiving additional induction therapy. Furthermore, the recent study of Cosio et al. in patients who received induction with thymoglobulin showed that in this group of patients, lower tacrolimus levels were not associated with higher tubulo-interstitial fibrosis and atrophy, but the contrary was true (23). Taken together, this could suggest that with the use of induction therapy high-tacrolimus levels are not needed to avoid progression of chronicity, even in higher risk recipients. However, due to the low number of patients in the induction group in our study, this conclusion is very premature.

The finding that higher exposure to tacrolimus was not associated with increased changes suggestive of CNIT (increase in arteriolar hyalinosis or new-onset arteriolar hyalinosis) does not imply that the histological entity CNIT is illusive. As has been shown by Nankivell et al. (8), the prevalence of arteriolar hyalinosis increases mainly after 1-year posttransplantation. It is thus possible that the current study could not demonstrate an association between tacrolimus levels and lesions suggestive of CNIT by its short follow-up time. The short follow-up time poses a limitation to this study, and the results should not be extrapolated beyond 1 year after transplantation. On the other hand, the lack of association between tacrolimus levels and increase in CNIT lesions could also suggest that other factors than systemic drug exposure contribute to tacrolimus nephrotoxicity. It could be argued that locally attained drug levels are more important than systemic drug exposure, as has been suggested by the tissue-specific distribution of CNI concentrations, and by the association of P-glycoprotein expression and CNIT (24,25). In addition, as the nephrotoxic effects of calcineurin inhibitors are at least partly mediated through variable degrees of oxidative stress, nuclear factor kappa-B dependent processes and other pathways (26,27), it can be expected that these factors influence the individual susceptibility for CNIT.

Study Limitations

The observational nature of this study, the short follow-up time and the relatively low number of patients pose major limitations to the clinical applicability of our finding that the chronic renal pathology at 12 months—and the increase of these chronic lesions in the preceding time interval—is independently associated with lower tacrolimus trough levels. Although this was previously shown in CsA-treated patients (9,10), this finding seems counter-intuitive and provocative, especially in the light of recent randomized studies demonstrating (in terms of renal graft function and acute rejection incidence; protocol biopsies were not performed) nonsuperiority of standard dose CsA versus low-dose CsA up to 3 years after transplantation (28,29), or superiority of low-dose tacrolimus over both normal-dose and low-dose CsA up to 1 year after transplantation (30). These trials used additional induction with daclizumab, and this could well be the main reason for the apparent discrepancies with our findings, as the majority of our patients did not receive additional induction therapy (see above). The fact that the subgroup of patients receiving additional induction therapy was too small to draw any conclusion in this subgroup, forms indeed a limitation to directly compare our results with the results of these large clinical trials. It may be clear that the findings of the current study regarding tacrolimus exposure need to be confirmed in a randomized trial comparing low versus standard dose tacrolimus, with close monitoring of the subclinical evolution of graft histology. We hope that our study will stimulate further debate and investigation on this topic. Finally, we recognize that the significant associations found in the current study do not prove a causal relationship between the clinical determinants and the histological outcome variables, nor does this study provide mechanistic explanations for the associations between clinical data and histology.

Conclusion

In conclusion, this study discerns two phases in the evolution of renal allograft histology in the first year after transplantation. In the first months, donor age and delayed graft function are associated with a less favorable subclinical evolution. In a later time frame, additional histological damage occurs independent of donor-related factors, but dependent on rejection phenomena and immune-mediated mechanisms. Tacrolimus doses or systemic exposure were not associated with lesions suggestive for calcineurin nephrotoxicity, suggesting that other factors determine susceptibility to tacrolimus nephrotoxicity.

Acknowledgments

We like to thank H. De Loor, M. Dekens, R. Eerdekens, A. Herelixcka and H. Wielandt for their continuous efforts in this study, and all colleagues of the Leuven Collaborative group for Renal Transplantation (LSGN) for their continuous support.

Funding sources: M. Naesens received a grant from the Fund for Scientific Research – Flanders (Belgium) (FWO).

Ancillary