Nephrol Dial Transplant (2006) 21: 3243–3251 doi:10.1093/ndt/gfl397 Advance Access publication 28 July 2006 Original Article

Conversion of ciclosporin A to tacrolimus in kidney transplant recipients with chronic allograft nephropathy

Sydney Chi-Wai Tang, Kwok Wah Chan, Colin Siu-On Tang, Man Fai Lam, Chung Ying Leung,

Kai Chung Tse, Chun Sang Li, Yiu Wing Ho, Matthew Kwok-Lung Tong, Kar Neng Lai and Downloaded from https://academic.oup.com/ndt/article/21/11/3243/1875343 by guest on 04 October 2021 Tak Mao Chan, for the Hong Kong Nephrology Study Groupy

Division of Nephrology, Department of Medicine, University of Hong Kong and Queen Mary Hospital, Hong Kong, China

Abstract patient number. Further studies with a larger cohort Background. Tacrolimus and ciclosporin might have are needed for validation. different effects on intra-renal fibrosis and allograft function in chronic allograft nephropathy (CAN). It is Keywords: chronic allograft nephropathy; difficult to predict the response to calcineurin inhibitor ciclosporin A; minimization; tacrolimus conversion minimization in patients with CAN. Methods. This prospective randomized study com- pared ciclosporin A (CsA)-to-tacrolimus conversion Introduction (group A, target tacrolimus trough level 6–8 ng/ml) vs CsA minimization (group B, target CsA trough level Despite improvements in immunosuppressive 80–100 ng/ml) with regard to efficacy and safety in protocols for kidney transplantation, chronic allograft patients with CAN and deteriorating allograft func- nephropathy (CAN) remains one of the most impor- tion. The primary efficacy endpoint was improvement tant causes of graft loss in the first post-transplant in the slope of inverse serum creatinine (1/SCr) vs decade [1]. CAN is characterized clinically by progres- time plot. sive graft failure associated with variable degrees of Results. There were 34 evaluable patients (n ¼ 16 in hypertension and proteinuria, and histologically by group A; n ¼ 18 in group B), with similar baseline non-specific chronic changes in the vascular (afferent characteristics. Both groups reached target drug levels arteriosclerosis), glomerular (glomerulosclerosis) and after a 3-month run-in period. Over the ensuing tubulointerstitial (tubular atrophy and interstitial 12 months, nine (56.3%) subjects in group A and fibrosis) compartments of the kidney [2]. Five years 10 (55.6%) in group B reached the primary end point after transplantation, up to two-thirds of allografts (P ¼ 0.968). Both groups showed considerable have been shown to demonstrate features of moderate improvement in the slope of 1/SCr vs time plot. to severe CAN [3]. The aetiology of CAN is not There was no significant difference in the slope well understood, and a variety of immunological between groups before and after intervention. Graft and nonimmunological factors, including histo- survival was 87% in group A and 100% in group B incompatibility, acute rejection, preservation injury, (P ¼ 0.121). Acute rejection was encountered in two donor status, hypertension, hyperlipidaemia and the group A subjects. There was no significant change or use of calcineurin inhibitors (CNI), notably ciclosporin difference in blood glucose, lipids, and blood pressure A (CsA), have been incriminated [1]. between groups. To date, there is no specific treatment for CAN. Conclusion. Our results suggest that in patients with Emerging evidence indicates that growth factors, CAN and deteriorating allograft function, CsA-to- such as transforming growth factor-b (TGF-b), are tacrolimus conversion or CsA minimization achieved critically important in both CAN and chronic CsA comparable efficacies in retarding the decline of graft toxicity, suggesting that these two entities share function. Such contention may be biased by the low common pathophysiological pathways, and chronic CsA nephrotoxicity may contribute significantly to CAN [4–7]. Because of this, several studies have investigated the role of an alternative immunosuppres- Correspondence and offprint requests to: Prof. T. M. Chan, Department of Medicine, Queen Mary Hospital, 102 Pokfulam sive agent, such as mycophenolate mofetil (MMF), and Road, Hong Kong SAR, China. Email: [email protected] found that the introduction of MMF with reduction yPlease refer to Appendix for participants in the Study Group. or withdrawal of CsA had a favourable outcome

ß The Author [2006]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please email: [email protected] 3244 S. C.-W. Tang et al. on post-transplantation graft function and survival dysfunction (de novo or recurrent disease, obstruction or [8,9]. However, such strategy, although safe in most transplant renal artery stenosis detected by Doppler ultra- instances, may be offset by the potentially increased sonography during the preceding 12 months), biopsy proven risk of acute rejection and graft loss [10,11], and it may acute rejection of Banff 97 grades I, II or III within 3 months not be entirely desirable to withdraw CNI altogether. prior to study entry, and patients who were receiving Tacrolimus is a newer CNI that is more potent tacrolimus or rapamycin before recruitment were excluded. than CsA in rejection prophylaxis [12], and has been Additional exclusion criteria were systemic infection, intro- associated with reduced intrarenal expression of duction of angiotensin converting enzyme inhibitor or TGF-b [13,14] and other profibrotic genes [15], and receptor blocker within 3 months before recruitment, receipt less allograft fibrosis [16] compared with patients of other solid organ transplant, concurrent participation treated with CsA. These observations suggest that in other investigational immunosuppressive protocol, and women lactating, pregnant or of childbearing potential not tacrolimus substitution for CsA may be beneficial for using, or who were unwilling to use a reliable contraceptive patients with documented CAN without compromising method during and for 6 weeks following the study. immunosuppressive efficacy. Downloaded from https://academic.oup.com/ndt/article/21/11/3243/1875343 by guest on 04 October 2021 In this study, we prospectively compared CsA-to- tacrolimus conversion with CsA minimization, with regard to renal function preservation in patients with Randomization and intervention protocol CAN and progressive loss of graft function. Eligible and consenting subjects were assigned by simple randomization using a computer-generated sequence into either the tacrolimus conversion (group A) or the CsA Methods minimization (group B) arm (Figure 1). Patients randomized to group A underwent a clean conversion to tacrolimus at a dose of 0.16 mg/kg/day in two divided doses 12 h after Study design and patients the last dose of CsA. The dose of tacrolimus was then This was a randomized, controlled, open-label, prospective titrated during a 3-month run-in phase to attain a 12-h study conducted in five major renal transplant centres in trough whole blood level of 6–8 ng/ml using a microparticle Hong Kong between July 2001 and June 2004. The study enzyme immunoassay on a Viva analyser (Dade Behring, protocol was approved by the Institutional Review Board UK). Other concomitant medications remained unchanged and Research Ethics Committee of the Hong Kong Hospital according to individual centre protocol. Patients randomized Authority, and all participating subjects gave written, to group B continued to receive CsA with dosages adjusted informed consent before study entry. Cadaveric or living- to maintain a 12-h trough blood level of 80–100 ng/ml related renal transplant recipients of either gender between using the Dade–Behring Emit Enzyme Immunoassay, and age 18 and 65, at least 12 months post-transplant and other medications according to centre protocol. All patients maintained on a CsA-based immunosuppressive regimen were followed-up every 2 weeks for the first month, then at with steady-state whole blood trough levels not exceeding 4-weekly intervals till 6 months, and then at 2-monthly 150 ng/ml and serum creatinine (SCr) between 100 and intervals till the study end. The total follow-up duration for 400 mmol/l, histologically proven CAN within the preceding both groups was 15 months. Pertinent clinical data including 12 months, and deteriorating graft renal function as body weight, blood pressure, urine dipstick test, complete evidenced by a negative slope of 1/SCr plotted against blood count, liver and renal function panels, blood glucose time, were eligible. A regression line plotted with at least six and 12-h trough blood levels of tacrolimus (for group A) or SCr values over the preceding 12 months had to demonstrate CsA (for group B), were recorded on each visit. Fasting a negative slope that had a significant P-value <0.05 and an blood glucose, haemoglobin A1C level, full lipid profile, adjusted R2 > 0.35. All biopsies were independently reviewed and 24-h urine for protein and creatinine were assayed and scored by a single pathologist in accordance with the every 2 months. All adverse events and medication log were Banff 97 criteria [2]. Patients with alternative causes of graft carefully documented throughout the study.

Group A: clean conversion to tacrolimus (n=16) Eligible subjects at least 12 months post-transplant Group B: CsA-minimizing protocol (n=18)

−12 or more 0 3 15 Month Screening for CAN and Run-in deteriorating graft function Phase

Randomization Attained Study target drug end level Fig. 1. Study design. CsA-to-tacrolimus for CAN 3245 Study end points in the overall distribution of the Banff 97 grades for CAN between groups (P ¼ 1.000). The underlying The primary end point was the rate of a significant renal diseases were IgA nephropathy in three, chronic improvement in graft function defined as a stabilization or reduction of SCr observed from 3–15 months of conversion glomerulonephritis in three, crescentic glomerulo- as reflected by a flattening or positive slope of the 1/SCr nephritis in one, and unknown in nine for group A; against time plot, absence of any major event necessitating IgA nephropathy in four, membranoproliferative premature termination of the randomized therapy, and the glomerulonephritis in one, focal segmental glomerulo- absence of graft loss. sclerosis in one, hypertensive nephrosclerosis in one, Secondary end points were graft and patient survival, chronic pyelonephritis in one, chronic glomerulo- incidence of acute rejection, SCr and calculated creatinine nephritis in two, and unknown in eight for group B. clearance, blood pressure, salient laboratory parameters, incidence of de novo post-transplant diabetes mellitus, use of medications for hypertension, diabetes mellitus and Immunosuppressive treatment hyperlipidaemia. The dose and trough blood level of CsA decreased Downloaded from https://academic.oup.com/ndt/article/21/11/3243/1875343 by guest on 04 October 2021 significantly in group B after the first 3 months, while the dose of prednisolone was comparable between Statistical analysis the two groups (Table 2). There was no significant Statistical analyses and sample size calculations were within-group difference in steroid dosage before and performed using SPSS for windows (version 12.0) and SAS after intervention. Most of the patients randomized to for windows (version 9.1). Our crude estimate indicated that CsA-to-tacrolimus conversion required a reduction the enrolment of 46 patients would achieve 80% power to in tacrolimus dose during titration, so that there was detect a 20% difference in SCr at 1 year between the two a significant reduction of tacrolimus dose for the groups with a two-sided significance level (a) of 0.05 using group as a whole during the first 3 months. The an independent two-sample t-test. Assuming a dropout rate final stable dose of tacrolimus was in the range of of 10%, the study aimed to enrol 50 patients. Due to the 0.066–0.08 mg/kg/day. stringent inclusion criteria, the final sample size achieved (n ¼ 34) resulted in a power of 67%. Data were presented as mean SD unless otherwise specified. Continuous variables Primary outcome between groups were compared using either unpaired t-test or Mann–Whitney U-test. Paired t-test or Wilcoxon signed Nine (56.3%) patients in group A and 10 (55.6%) rank test was used for within group comparison where patients in group B (P ¼ 0.968) showed a significant appropriate. Comparisons of categorical variables were improvement in 1/SCr vs time slope, indicating a performed using either w2 test or Fisher’s exact test. Linear reduction in the rate of renal function deterioration, regression analysis was used to estimate the slope of the following intervention (Figure 2). There was a sig- 1/SCr vs time plot for individual subjects. SCr values within nificant reduction in 1/SCr vs time slope in group A the first 3 months after randomization were not included in (overall trend for the whole group before and after the calculation of slope of 1/SCr vs time in view of the blood intervention, 0.123 vs 0.050 l/mmol/month, level titration during this period. Differences in the slopes P ¼ 0.029) (Figure 3). The corresponding values in of 1/SCr vs time before (calculated from SCr values in the group B were 0.090 and 0.045 l/mmol/month 12 months before randomization) and after intervention (P ¼ 0.096). Between-group comparison showed (calculated from SCr values over the 3–15 months interval that their difference in 1/SCr vs time slope was following randomization) and between the two treatment not statistically significant both before (difference groups, were sought [17]. Mixed linear model was used in in slope of mixed linear model regression the calculation of least squares means of SCr and in the 0.033 l/mmol/month, P ¼ 0.196) and after inter- comparison of the trends of 1/SCr vs time plots between vention (difference in slope of mixed linear model the two groups. Treatment group and time were treated as regression 0.010 l/mmol/month, P ¼ 0.724). fixed effects in the mixed linear model analysis. Akaike’s information criterion was used to assess model fitness. Two-tailed P-values of <0.05 were considered statistically Secondary outcomes significant. Changes in other salient clinical and laboratory parameters are listed in Table 3. Both treatment Results groups showed a significant and sustained reduction in systolic blood pressure compared with baseline values within the first 6 months. Proteinuria and Baseline characteristics creatinine clearance did not change significantly and All participating subjects were ethnic Chinese. The remained similar between the two groups. Triglyceride baseline demographic and clinical characteristics of level in group B was significantly lower after 6 months, the two groups were well matched except for a higher but the lipid profile, blood glucose and uric acid levels proportion of subjects with a history of allograft did not differ between the two groups throughout the rejection and concomitant use of MMF or azathio- study. The reduction in uric acid level in both groups prine in group A (Table 1). There was no difference did not reach statistical significance. None of the 3246 S. C.-W. Tang et al. Table 1. Baseline demographic and clinical characteristics of study subjects

CsA-to-tacrolimus Group A CsA minimization Group B P (n ¼ 16) (n ¼ 18)

Age (years) at randomization 48.5 10.1 45.2 10.8 0.367 Gender (M:F) 11:5 10:8 0.429 Body mass index (kg/m2) 22.9 3.1 22.7 3.0 0.953 Time from transplant to randomization (months) 92 49 99 52 0.685 Nature of transplants Cadaveric:living-related 13:3 13:5 0.536 First:second graft 15:1 17:1 0.894 Banff 97 score (2) for CANa 1.000b Grade I (a) 5 (31.3%) 6 (33.3%) 1.000 Grade II (a) 7 (43.8%) 7 (38.9%) 1.000

Grade III (a) 4 (25.0%) 5 (27.8%) 1.000 Downloaded from https://academic.oup.com/ndt/article/21/11/3243/1875343 by guest on 04 October 2021 Other immunosuppressive agents MMF 3 (18.8%) 1 (5.6%) 0.233 Azathioprine 5 (31.3%) 2 (11.1%) 0.214 Either 8 (50.0%) 3 (16.7%) 0.038 Metabolic disorders Diabetes mellitus 0 3 (16.7%) 0.087 Hyperlipidaemia on statin 3 (18.8%) 2 (11.1%) 0.648 History of previous acute rejection 13 (81.3%) 6 (33.3%) 0.002 1 episode 7 (43.8%) 4 (22.2%) 0.274 2 episodes 4 (25%) 2 (11.1%) 0.387 3 episodes 2 (12.5%) 0 0.214 SCr (mmol/l) Mean 220 51.9 213 78.3 0.293 Range 120–315 136–388 Creatinine clearance (ml/min) Mean 36.2 13.0 41.6 20.3 0.771 Range 18.7–58.5 21.4–88.2 aNone of the subjects had features of glomerular and vascular lesions suggestive of chronic/recurrent rejection. Thus, all subjects had Banff sub-grade (a). bOverall P-value for comparison of grade distribution between groups by exact test. CAN, chronic allograft nephropathy; MMF, mycophenolate mofetil; SCr, serum creatinine.

Table 2. Dosages and trough blood levels of immunosuppressants at baseline, 3, 6 and 12 months

CsA-to-tacrolimus Group A (n ¼ 16) CsA minimization Group B (n ¼ 18) P

CsA dose (mg/day) Baseline 135.7 31.8 145.6 53.2 0.820 3 months – 120.9 34.3 0.047b 6 months – 119.1 31.3 0.045b 12 months – 122.1 38.4 0.010b CsA trough (ng/ml)a Baseline 114.8 23.5 132.3 42.7 0.300 3 months – 121.0 40.2 0.078b 6 months – 90.5 28.5 <0.001b 12 months – 93.9 33.7 0.040b Tacrolimus dose (mg/day) Baseline 8.0 2.3 – 3 months 4.5 2.0 – 0.002b 6 months 4.8 2.1 – 0.017b 12 months 4.1 1.8 – 0.027b Tacrolimus trough (ng/ml)a Baselinex 14.0 7.8 – 3 months 7.9 2.7 – 0.002b 6 months 8.0 3.3 – 0.009b 12 months 6.8 1.3 – 0.003b Prednisolone dose (mg/day) Baseline 8.0 2.5 7.0 1.7 0.235 3 months 7.5 1.7 6.9 1.6 0.333 6 months 7.0 1.6 6.9 1.6 0.812 12 months 6.9 1.7 6.7 1.4 0.812 aMeasured at 12 h post-dose. bWithin-group comparison vs baseline value. xMeasured 1 week after the commencement of tacrolimus. CsA-to-tacrolimus for CAN 3247

2 2 CsA-to-FK CsA ) ) 3 1 3

10 1 10 × ×

0 0 mol/year mol/year µ − µ 1 −1 time (L/ −2 time (L/ −2 vs vs

− 3 −3

−4 −4 Slope of 1/SCr Downloaded from https://academic.oup.com/ndt/article/21/11/3243/1875343 by guest on 04 October 2021 Slope of 1/SCr − 5 −5 Before After Before After

2 2 CsA-to-FK CsA ) ) 3 3 0 0 10 10 × ×

−2 −2 mol/year mol/year

µ * µ −4 −4

time (L/ −6 time (L/ −6 vs vs

−8 −8

−10 −10 Slope of 1/SCr Slope of 1/SCr * −12 −12 Before After Before After Fig. 2. Slope of 1/SCr vs time plot in individual patients before and after intervention. Upper panels include patients who showed a significant improvement after intervention. Lower panels include the remaining patients in whom the slope of 1/SCr vs time either did not improve or deteriorated. *Indicate the two patients who developed end-stage renal failure during follow-up. CsA-to-FK: ciclosporin conversion to tacrolimus (group A); CsA: ciclosporin minimization (group B).

7 Run-in subjects developed de novo diabetes mellitus, and there phase 6 was no change in the requirement for anti-hypertensive or lipid-lowering medications. 3 5 Graft survival in the study period was 87% in group 10 × 4 A and 100% in group B (P ¼ 0.121, Figure 4). Two

mol/l) group A patients reached end-stage renal failure, one µ 3 due to progression of CAN and the other one followed

1/SCr ( 2 an episode of delayed acute rejection. Acute rejection CsA-to-FK occurred in two patients, both from group A. One 1 CsA patient developed acute rejection 2 months after 0 entering this study, which was related to suboptimal −12 −9 −6 −303691215 tacrolimus blood levels. His baseline SCr was Time after randomization (months) 300 mmol/l. Despite treatment with pulse corticosteroid Fig. 3. Mixed linear model regression analysis of 1/SCr. Time zero he became dialysis dependent four weeks later. refers to the point of randomization. There was a significant reduction in overall trend of renal deterioration for patients in Another group A subject developed Banff 97 grade group A following intervention (0.123 vs 0.050 l/mmol/month IA acute rejection one week after entering this study, before and after intervention, respectively, P ¼ 0.029), which was not when the trough tacrolimus blood level was 8.9 ng/ml. observed in group B (0.090 and 0.045 l/mmol/month before and He responded to pulse corticosteroid treatment and after intervention, respectively, P ¼ 0.096). Between-group difference was not significant before (P ¼ 0.196) or after intervention SCr returned to baseline level. The incidence of adverse (P ¼ 0.724). CsA-to-FK: ciclosporin conversion to tacrolimus events did not differ between the two groups (Table 4). (group A); CsA: ciclosporin minimization (group B). Four patients in group A and one in group B 3248 S. C.-W. Tang et al. Table 3. Serial profile of blood pressure, renal function, and metabolic parameters

CsA-to-tacrolimus Group A (n ¼ 16) CsA minimization Group B (n ¼ 18) P

Systolic blood pressure (mmHg) Baseline 135 17 134 15 0.629 6 months 126 13* 123 16** 0.427 12 months 130 21 127 17 0.575 Diastolic blood pressure (mmHg) Baseline 77 10 76 6 0.489 6 months 75 872 12 0.438 12 months 77 10 76 12 0.794 No. of antihypertensive drugs Baseline 2.0 1.03 2.4 1.14 0.159 12 months 2.5 1.13 2.8 1.25 0.141

Creatinine clearance (ml/min) Downloaded from https://academic.oup.com/ndt/article/21/11/3243/1875343 by guest on 04 October 2021 Baseline 36.2 13.0 41.6 20.3 0.771 12 months 32.9 11.1 39.8 20.2 0.534 Urine protein excretion (g/24 h) Baseline 1.53 2.09 0.60 0.84 0.114 12 months 1.84 3.47 0.45 0.60 0.191 Total cholesterol (mmol/l) Baseline 5.6 0.7 5.7 1.5 0.752 6 months 5.2 1.2 5.4 0.9 0.154 12 months 5.1 0.8 5.3 1.0 0.356 Serum triglyceride (mmol/l) Baseline 2.27 0.87 2.21 1.44 0.285 6 months 1.92 0.72 1.70 0.81*** 0.217 12 months 2.14 0.89 1.53 0.58 0.114 No. of subjects receiving statin Baseline 3 2 0.648 12 months 2 4 0.634 Serum uric acid (mmol/l) Baseline 559 129 564 208 0.970 12 months 513 97 489 128 0.698 Fasting glucose (mmol/l) Baseline 5.1 1.2 5.0 0.9 0.956 12 months 5.4 1.1 4.9 1.0 0.180 Haemoglobin A1C(%) Baseline 5.9 0.5 6.1 0.8 0.406 12 months 6.3 0.3 5.9 0.5 0.291

*P ¼ 0.035, **P ¼ 0.011, ***P ¼ 0.049 vs corresponding baseline value. All other within-group differences between 6/12-month and baseline values were statistically insignificant.

100 CsA Discussion

CsA-to-FK 80 The results from this study showed that in renal transplant recipients of CsA-based immunosuppres- 60 sion with deteriorating allograft function due to CAN, conversion to tacrolimus or reduction of CsA alone

40 are equally effective in improving the subsequent

Graft survival (%) course of CAN, as evidenced by a significant flattening of the negatively sloping 1/SCr vs time plot in the 20 ensuing 12 months after therapeutic intervention. This P=0.121 beneficial effect was observed in over half of the 0 patients in each group. When the overall trend for the 0 2 4 6 8 10 12 14 16 whole group was analysed, a significant improvement Time after randomization (months) was observed only in the CsA-to-tacrolimus group Fig. 4. Graft survival rates. CsA-to-FK: ciclosporin conversion to (group A) but not with CsA minimization. This tacrolimus (group A); CsA: ciclosporin minimization (group B). discrepancy could be related to the more negative slope in group A subjects at baseline compared with developed infection (P ¼ 0.164). There was no mortal- group B, although the between-group difference did ity. Apart from the two patients who reached end-stage not reach statistical significance both before and after renal failure, none required premature discontinuation intervention. These findings bear clinical importance. of study. Firstly, gradual minimization of CsA is both safe CsA-to-tacrolimus for CAN 3249 Table 4. Adverse events

CsA-to-tacrolimus Group A (n ¼ 16) CsA minimization Group B (n ¼ 18)

Acute rejection 2 (13%)a 0 Infection Urinary tract infection 1 (6%) 0 Gastroenteritis 0 1 (6%) Herpes zoster 1 (6%) 0 Cytomegalovirus disease 0 0 Fever without positive culture 2 (13%) 0 Gouty arthritis 4 (25%) 1 (6%) Tremor 1 (6%) 0 Hot flushes 1 (6%) 1 (6%) Pruritus and skin rashes 1 (6%) 1 (6%) Downloaded from https://academic.oup.com/ndt/article/21/11/3243/1875343 by guest on 04 October 2021 aOne episode led to end-stage renal failure, another resolved after pulse methylprednisolone treatment. and effective without the need to introduce another levels is supported by a recent observational study [24] immunosuppressive agent. None of the subjects in in which conversion of CsA to tacrolimus with trough the CsA arm experienced acute rejection after blood levels of 6–7 ng/ml in 133 renal transplant recipients levels of CsA were reduced by an average of 29% was adequate and not associated with increased at 12 months. The final drug levels were similar to rejection risk after 4 years of follow-up. The immediate those in a recent study [18] of 13 subjects with CAN advantages of low-dose tacrolimus include reduced who benefited from a 24% dose reduction in CsA. risk of side effects, notably new-onset diabetes which However, the investigators in this study had also is dose-dependent [12], and lowering of drug budget. introduced MMF prior to CsA reduction. Our results Indeed, the incidence of de novo diabetes approached suggest that this may not be necessary in all patients, 8.5% in 107 patients who underwent tacrolimus which has obvious financial implications. conversion when the drug level was aimed high [28]. Secondly, tacrolimus conversion or CsA reduction On the other hand, patients converted to tacrolimus yielded comparable profiles of SCr regression and, from CsA in the present study were at no greater more importantly, obviated the potentially heightened risk for developing diabetes or glucose intolerance, risk of acute rejection observed with CNI-withdrawing which concurs with observations made in the protocols. Although the addition of MMF followed by Caucasian population employing lower drug targets complete CNI withdrawal has been reported to achieve [18,29,30]. Furthermore, the long term metabolic and success for CAN [8,19], the risk of acute rejection is organ-specific benefits of reduced cumulative exposure significantly higher following CsA withdrawal either in to CNI may be even more far-reaching. In addition stable patients receiving MMF [20], or during the early to diabetes, lipid profiles were consistently reported to post-transplant period in patients receiving regimens be improved after CsA was switched to tacrolimus that contain MMF [21] or sirolimus [22]. Indeed, [23,24]. However, we and others [26] have not a meta-analysis of 10 clinical trials including data from reproduced this phenomenon. The discrepancy could over 1000 subjects demonstrated an 11% increase in be related to the confounding effect of variable doses the odds of acute rejection after CsA withdrawal [11]. of corticosteroid in different series. Thus, while the safety of complete CNI withdrawal One drawback of this study is the low patient is still debatable, a minimization approach may be number which could bias our conclusions. This arises a viable alternative option that alters the course of from the stringent inclusion criteria that resulted CAN without compromising safety. in a slow take-on rate. In addition, randomization Thirdly, although there have been several reports of low patient numbers resulted in an imbalance in favouring the conversion of CsA to tacrolimus concomitant immunosuppressants and historical rejec- for CAN [18,23–27], these studies were often uncon- tion rates at baseline. Nevertheless, such differences, trolled or retrospective in nature. The recommended even if real, would only be in favour of a stabilizing tacrolimus dose and target trough blood levels effect of CsA-FK506 conversion. More importantly, from two recent randomized controlled trials were there was no difference in the Banff 97 score 0.15 mg/kg/day and 5–15 ng/ml [23], and 0.10 mg/kg/ distribution between groups. Herein, around 55% of day and 5–10 ng/ml [18], respectively, while the subjects randomized to either treatment arm reached corresponding values in 2 non-randomized trials were primary endpoint. To our knowledge, there is only even higher at 0.15 mg/kg/day and 10–15 ng/ml [27,28], one published study that also employs 1/SCr vs respectively. Here, we showed that much lower time trend as the primary end point in tacrolimus- doses of 0.066–0.08 mg/kg/day and trough levels converted subjects [18], and the response rates were within 6–8 ng/ml produced comparable results without slightly higher, being 62% (8 of 13 subjects by compromising safety. The feasibility of such low blood intention-to-treat analysis) in the conversion arm and 3250 S. C.-W. Tang et al. 69% (9 of 13 subjects) in the CsA reduction plus MMF Acknowledgements. This study was partially supported by a grant arm. On the other hand, absolute SCr values were used from Fujisawa HK Ltd. The authors thank Ms Emily W. L. Liu for accurately recording all clinical data from the participating centres. as the primary outcome measure in two recently published trials of tacrolimus conversion in which SCr decreased after 2 years [23] and 4 years [24] of Conflict of interest statement. None declared. conversion. Here, we only demonstrated a slower pace of SCr rise rather than a decline in the absolute SCr value. Such discrepancy may be partly due to the References exclusion of subjects who lost their grafts in those two studies, and partly related to the shorter follow-up 1. Paul LC. Chronic allograft nephropathy: An update. Kidney Int 1999; 56: 783–793 duration in this study. Nevertheless, a potential 2. Racusen LC, Solez K, Colvin RB et al. The Banff 97 working disadvantage of using SCr alone as the treatment end classification of renal allograft pathology. Kidney Int 1999; 55: point is that it only reflects graft function at a pre- 713–723 selected time point rather than its true evolution over 3. Nankivell BJ, Borrows RJ, Fung CL, O’Connell PJ, Allen RD, Downloaded from https://academic.oup.com/ndt/article/21/11/3243/1875343 by guest on 04 October 2021 time. We believe that it is clinically pertinent to Chapman JR. The natural history of chronic allograft nephro- compare the rate of renal function deterioration pathy. N Engl J Med 2003; 349: 2326–2333 4. Pascual M, Swinford RD, Ingelfinger JR, Williams WW, in individual patients before and after intervention, Cosimi AB, Tolkoff-Rubin N. Chronic rejection and chronic since SCr value can be affected by a change in muscle cyclosporin toxicity in renal allografts. Immunol Today 1998; 19: mass, and the rate of progression of renal failure varies 514–519 between patients. 5. Shin GT, Khanna A, Ding R et al. In vivo expression of Graft survival and the pattern of adverse events transforming growth factor-beta1 in humans: stimulation by were comparable between the two treatment groups cyclosporine. Transplantation 1998; 65: 313–318 although two patients developed acute rejection in 6. Pankewycz OG, Miao L, Isaacs R et al. Increased renal tubular expression of transforming growth factor beta in human group A. One patient developed delayed acute rejec- allografts correlates with cyclosporine toxicity. Kidney Int tion that was probably precipitated by a short period 1996; 50: 1634–1640 of suboptimal tacrolimus blood levels. Thus, from a 7. Cuhaci B, Kumar MS, Bloom RD et al. Transforming growth practical viewpoint, a minimization protocol is more factor-beta levels in human allograft chronic fibrosis correlate convenient than a switch in CNI, since the latter with rate of decline in renal function. Transplantation 1999; 68: entails starting with an empirical dose that may not 785–790 always be adequate. This does not mean our empirical 8. Dudley C, Pohanka E, Riad H et al. Mycophenolate mofetil substitution for cyclosporine a in renal transplant recipients starting dose of tacrolimus at 0.16 mg/kg/day was with chronic progressive allograft dysfunction: the ‘‘creeping inadequate, since the majority of patients required creatinine’’ study. Transplantation 2005; 79: 466–475 a subsequent downward adjustment so that 9. Weir MR, Ward MT, Blahut SA et al. Long-term impact of by 12 months there was a 49% dose reduction. discontinued or reduced calcineurin inhibitor in patients with Another patient experienced rejection just one chronic allograft nephropathy. Kidney Int 2001; 59: 1567–1573 week after entering the study, which was likely to 10. Schrama YC, Joles JA, van Tol A, Boer P, Koomans HA, be related to the preceding low CsA blood levels in Hene RJ. Conversion to mycophenolate mofetil in conjunction with stepwise withdrawal of cyclosporine in stable renal view of the concomitant tacrolimus level of 8.9 ng/ml. transplant recipients. Transplantation 2000; 69: 376–383 In a series that compared 126 patients who 11. Kasiske BL, Chakkera HA, Louis TA, Ma JZ. A meta-analysis underwent CsA-to-tacrolimus conversion with 60 of immunosuppression withdrawal trials in renal transplanta- patients who remained on CsA, the incidence of tion. J Am Soc Nephrol 2000; 11: 1910–1917 acute rejection in the first 2 years was 4.8 and 5%, 12. Knoll GA, Bell RC. Tacrolimus versus cyclosporin for respectively [23]. immunosuppression in renal transplantation: meta-analysis of Apart from CsA minimization or conversion to randomised trials. Br Med J 1999; 318: 1104–1107 13. Matl I, Viklicky O, Voska L, Lodererova A, Vitko S. The effect tacrolimus or MMF, there is an emerging trend of different immunosuppressive regimens on TGF-beta1 expres- to combine CNI minimization or withdrawal with sion in kidney transplant patients. Transpl Int 2005; 18: 668–671 mammalian target of rapamycin (mTOR) inhibitors 14. Mohamed MA, Robertson H, Booth TA, Balupuri S, Kirby JA, such as sirolimus or everolimus [31]. Absence of Talbot D. TGF-beta expression in renal transplant biopsies: nephrotoxicity is a distinct advantage of mTOR a comparative study between cyclosporin-A and tacrolimus. inhibitors, but they were not available at the inception Transplantation 2000; 69: 1002–1005 of this study. In a recent report of 43 patients with 15. Bicknell GR, Williams ST, Shaw JA, Pringle JH, Furness PN, Nicholson ML. Differential effects of cyclosporin and tacroli- CAN, conversion from CNI to sirolimus was asso- mus on the expression of fibrosis-associated genes in isolated ciated with improved renal function, yet up to a third glomeruli from renal transplants. Br J Surg 2000; 87: 1569–1575 of the subjects developed overt proteinuria [32]. The 16. Murphy GJ, Waller JR, Sandford RS, Furness PN, long-term safety and efficacy of this approach requires Nicholson ML. Randomized clinical trial of the effect of further validation. microemulsion cyclosporin and tacrolimus on renal allograft In summary, this study indicates comparable fibrosis. Br J Surg 2003; 90: 680–686 17. Zar JH. Comparing simple linear regression equations. safety and efficacy of CsA-to-tacrolimus conversion In: Zar JH, Ryu Y, Snavely SL, eds. Biostatistical Analysis, and CsA minimization in patients with CAN and Upper Saddle River, NJ, USA: Prentice Hall, 1999; 360–376 deteriorating graft function. Its validity remains to be 18. Stoves J, Newstead CG, Baczkowski AJ, Owens G, Paraoan M, confirmed in a larger cohort of subjects. Hammad AQ. A randomized controlled trial of CsA-to-tacrolimus for CAN 3251 immunosuppression conversion for the treatment of chronic 28. Yagmurdur MC, Sevmis S, Emiroglu R, Moray G, Bilgin N, allograft nephropathy. Nephrol Dial Transplant 2004; 19: Haberal M. Tacrolimus conversion in kidney transplant 2113–2120 recipients: analysis of 107 patients. Transplant Proc 2004; 36: 19. Suwelack B, Gerhardt U, Hohage H. Withdrawal of cyclospor- 144–147 ine or tacrolimus after addition of mycophenolate mofetil in 29. First MR, Gerber DA, Hariharan S, Kaufman DB, Shapiro R. patients with chronic allograft nephropathy. Am J Transplant Posttransplant diabetes mellitus in kidney allograft recipients: 2004; 4: 655–662 incidence, risk factors, and management. Transplantation 2002; 20. Abramowicz D, Manas D, Lao M et al. Cyclosporine with- 73: 379–386 drawal from a mycophenolate mofetil-containing immuno- 30. Margreiter R. Efficacy and safety of tacrolimus compared suppressive regimen in stable kidney transplant recipients: with ciclosporin microemulsion in renal transplantation: a randomized, controlled study. Transplantation 2002; 74: a randomised multicentre study. Lancet 2002; 359: 741–746 1725–1734 31. Chapman JR, O’Connell PJ, Nankivell BJ. Chronic renal 21. Smak Gregoor PJ, de Sevaux RG, Ligtenberg G et al. allograft dysfunction. J Am Soc Nephrol 2005; 16: 3015–3026 Withdrawal of cyclosporine or prednisone six months after 32. Bumbea V, Kamar N, Ribes D et al. Long-term results in renal kidney transplantation in patients on triple drug therapy: transplant patients with allograft dysfunction after switching

a randomized, prospective, multicenter study. J Am Soc from calcineurin inhibitors to sirolimus. Nephrol Dial Transplant Downloaded from https://academic.oup.com/ndt/article/21/11/3243/1875343 by guest on 04 October 2021 Nephrol 2002; 13: 1365–1373 2005; 20: 2517–2523 22. Johnson RW, Kreis H, Oberbauer R, Brattstrom C, Claesson K, Eris J. Sirolimus allows early cyclosporine withdrawal in renal Received for publication: 17.1.06 transplantation resulting in improved renal function and lower Accepted in revised form: 6.6.06 blood pressure. Transplantation 2001; 72: 777–786 23. Waid T. Tacrolimus as secondary intervention vs cyclosporine continuation in patients at risk for chronic renal allograft failure. Clin Transplant 2005; 19: 573–580 Appendix 24. Cantarovich D, Renou M, Megnigbeto A et al. Switching from cyclosporine to tacrolimus in patients with chronic Membership of the Hong Kong Nephrology Study transplant dysfunction or cyclosporine-induced adverse events. Group includes: Queen Mary Hospital—Sydney CW Transplantation 2005; 79: 72–78 25. Hohage H, Welling U, Zeh M, Gerhardt U, Suwelack B. Tang, Sing Leung Lui, Angela YM Wang, Kwok Wah Switching immunosuppression from cyclosporine to tacrolimus Chan, Colin SO Tang, Man Fai Lam, Kai Chung Tse, improves long-term kidney function: a 6-year study. Transplant Terence PS Yip, Kar Neng Lai, Tak Mao Chan; Proc 2005; 37: 1898–1899 United Christian Hospital—Anthony Tang, Yiu Wing 26. Lee WC, Lian JD, Wu MJ, Cheng CH, Chen CH, Shu KH. Ho, Chung Ying Leung; Princess Margaret Hospital— Long-term beneficial effect of tacrolimus conversion on renal Ernest WK Tsang, Matthew KL Tong; Queen transplant recipients. Ren Fail 2005; 27: 501–506 Elizabeth Hospital—Francis KM Wong, Ka Foon 27. Morris-Stiff GJ, Baboolal K, Dunstan F, Jurewicz WA. Conversion from cyclosporin (Neoral) to tacrolimus (Prograf) Chau, Chun Sang Li; Kwong Wah Hospital—Kin in renal allograft recipients with chronic graft nephropathy: Yee Lo, Siu Ka Mak, Andrew KM Wong; Tuen Mun results of an observational study. Transpl Int 1999; 12: 288–292 Hospital—Tze Hoi Kwan, Tak Cheung Au.