Mycophenolic acid pharmacokinetics and related outcomes early after renal transplant

Abstract

Aims

The pharmacokinetics of mycophenolic acid and its glucuronide are complex. This study investigated the pharmacokinetics, pharmacodynamics and protein binding of mycophenolic acid and its glucuronide metabolite, early post-transplant in renal allograft recipients.

Methods

Forty-two de novo renal transplant recipients receiving mycophenolate mofetil and concomitant cyclosporin (n = 32) or tacrolimus (n = 10) participated in the study. Blood samples were taken on day 5 post-transplant for measurement of free and total concentrations of mycophenolic acid, mycophenolic acid glucuronide and relevant biochemistry. Associations between free fraction and biochemistry were investigated. Free and total 6-h area under the concentration–time curve (AUC₀₋₆) of mycophenolic acid was assessed relative to clinical outcomes in the first month post-transplant.

Results

Kinetic variability of free and total mycophenolic acid and its glucuronide was greater in patients on cyclosporin (12- to 18-fold variation) than on tacrolimus (four- to fivefold) cotherapy. Cyclosporin-treated patients also had significantly lower predose total mycophenolic acid concentrations than tacrolimus-treated patients (median 0.8 mg l⁻¹ and 1.6 mg l⁻¹, respectively, P = 0.002). Mycophenolic acid glucuronide predose concentration correlated positively with mycophenolic acid glucuronide AUC₀₋₆ (r > 0.95). Mycophenolic acid free fraction varied 11-fold, from 1.6% to 18.3%, whilst the glucuronide free fraction varied threefold, from 17.4% to 54.1%. Urea and creatinine concentrations correlated positively (r > 0.46), whilst albumin correlated negatively (r = −0.54) with free fraction of mycophenolic acid. Similar relationships were found for the free fraction of mycophenolic acid glucuronide. Mycophenolic acid free fraction was on average 70% higher in patients with albumin concentrations below a specified albumin cut-off concentration of 31 g l⁻¹[free fraction = 7 ± 4% for lower albumin and 4 ± 3% for higher albumin, respectively; P = 0.001; 95% confidence interval (CI) for the difference 1.9, 4.2]. Neither free nor total mycophenolic acid AUC₀₋₆ was related to rejection (P > 0.07). Free AUC₀₋₆ was significantly higher in those patients with thrombocytopenic, leukopenic and/or infectious outcomes than in those without (mean ± SD 1.9 ± 0.3 mg h⁻¹ l⁻¹ and 1.1 ± 0.1 mg h⁻¹ l⁻¹, P = 0.0043; 95% CI for the difference 0.3, 1.4).

Conclusions

The marked variability in mycophenolic acid/glucuronide pharmacokinetics occurring early post-transplant during the current study was greater in cyclosporin (12–18-fold) than in tacrolimus (four- to fivefold) treated patients. Concomitant cyclosporin was associated with total mycophenolic acid concentrations approximately half that of tacrolimus. Patients with marked renal impairment had the highest free fractions reported to date. The exposure to unbound mycophenolic acid was significantly related to infections and haematological toxicity.

Introduction

Mycophenolate mofetil (MMF), the prodrug of mycophenolic acid, is an immunosuppressant coadministered with corticosteroids and calcineurin inhibitors for the prevention of renal allograft rejection [1–5]. Mycophenolic acid (MPA) is metabolized to an inactive phenolic glucuronide (MPAG) that either undergoes enterohepatic cycling or is eliminated in the urine [6].

The pharmacokinetics of MPA in adult kidney transplant recipients is extremely complex with a 10-fold range in area under the concentration–time curve (AUC) for a given dose [7–11]. MPA AUC has predictive value for the risk for acute rejection [12]. A large multicentre study suggested that there is an inverse relationship between the MPA AUC and biopsy-proven rejection, with decreasing rejection rates as mean AUC values increase from 16.1 to 60.6 mg h⁻¹ l⁻¹ [13].

The most common adverse drug reactions are gastrointestinal upset and haematological disorders such as leukopenia [2, 14, 15]. In contrast to the data relating MPA pharmacokinetics to clinical efficacy, there is no consensus in the literature as to whether MMF dose [13], predose plasma MPA concentration [16], or 12-h MPA AUC [17, 18] correlate best with adverse effects.

MPA is 97–98% protein bound and it is the unbound, or free MPA that is pharmacologically active [19]. Multiple factors, including hypoalbuminaemia, uraemia and accumulated MPAG influence the protein binding of MPA in renal failure, alter free fraction in vivo and have consequences for its pharmacokinetics and pharmacodynamics [7, 19, 20]. Free and total concentration profiles of both MPA and MPAG have been inadequately characterized, especially in the early period after transplant [7, 21, 22].

The aim of this study was to investigate the pharmacokinetics of both total and free MPA and its glucuronide metabolite in an adult renal population early after transplantation and relate free concentrations of MPA and MPAG, with outcomes in the first month of therapy.

Methods

Study population and design

All adult (≥18 years old) patients receiving a renal transplant between June 2002 and April 2003 at the Princess Alexandra Hospital who were to receive MMF were invited to participate. Informed consent was obtained and the Princess Alexandra Hospital Research Ethics Committee approved the study protocol (No. 59/02). Pregnant females were excluded. All patients received oral MMF (CellCept®, 1 g b.i.d., commenced immediately preoperatively) with cyclosporin (Neoral®, 4 mg kg⁻¹ body weight b.i.d.; n = 32) or tacrolimus (Prograf®, 0.1 mg kg⁻¹ body weight; n = 10) and simulect (Basiliximab®; 20 mg days 0 and 4), diltiazem (240 mg daily throughout the study, for augmentation of calcineurin inhibitor exposure) and prednisolone (0.3 mg kg⁻¹ daily). The assignment of patients to cyclosporin or tacrolimus groups was dependent on the clinical transplant team. Tacrolimus was more likely to be used in those patients with a higher immunological risk as assessed by a panel-reactive antibody status. Patients were studied on day 5 post-transplant after MPA concentrations had reached steady state. Demographic data included age, sex, race, weight and height, body mass index, diabetic status, and cause of end-stage renal failure, if relevant. Transplantation details recorded included the transplant number, donor type, ischaemic time, need for continuous ambulatory peritoneal dialysis (CAPD) pretransplant, human leucocyte antigen haplotype mismatch (HLAM) number, panel-reactive antibodies (PRA) and delayed graft function (DGF – identified by the need for dialysis after transplant). HLAM and PRA were measured according to routine standard methods approved by the National Organ Matching System (NOMS) [23].

Routine pathology and therapeutic drug monitoring

Haemoglobin, white cell and platelet count, concentrations of plasma creatinine, urea, urate, total protein, albumin, globulin, bilirubin, bicarbonate and liver enzymes (ALT, AST, ALP, GGT) were recorded daily post-transplant. Creatinine clearance (CC) was estimated using the Cockroft–Gault formula [24] using lean body mass estimations.

Concentrations of cyclosporin or tacrolimus were measured daily for the first month post-transplant using previously published high-performance liquid chromatography (HPLC) [25] and HPLC-MS [26] methods, respectively. Cyclosporin dose was adjusted to obtain a target 2-h postdose whole-blood concentration of 1300–1700 µg l⁻¹ and tacrolimus dose was adjusted to obtain a target trough concentration of 10–20 µg l⁻¹.

MPA and MPAG analysis

Blood samples were collected into Vacutainer® tubes containing EDTA at predose and 1, 3 and 6 h after MMF dosage. This limited sampling regimen was based on our previously published model which demonstrated a correlation between abbreviated and full AUC of r² = 0.836 [27]. Blood samples were centrifuged for 10 min at 860 × g, and the plasma removed. Ultrafiltrate was prepared from 1 ml of plasma by ultracentrifugation at 3000 × g in a Beckman fixed rotor centrifuge (20 min, 20 °C).

Total concentrations of MPA and MPAG, and free concentrations of MPAG, were determined in plasma, and ultrafiltrate, respectively, using an HPLC-UV method [28]. Free MPA concentrations were measured in ultrafiltrate using a validated HPLC-MS-MS method [29].

The truncated areas under the concentration–time curve from 0 to 6 h (AUC₀₋₆) for free and total MPA were calculated using the linear trapezoidal rule. The MPA free fraction was calculated for each patient by dividing the AUC₀₋₆ free by the AUC₀₋₆ total (× 100%).

Pharmacodynamic outcomes

Information on biopsy-proven rejection according to Banff (1997) criteria and on MMF toxicity (principally gastrointestinal, haematological and infectious adverse events) in the first month after renal transplantation was collected from medical records. Gastrointestinal adverse effects were defined as the onset of nausea, vomiting, abdominal cramping or diarrhoea (passage of >200 ml day⁻¹ watery stool). Haematological adverse effects were defined as the development of leukopenia (WCC <4 × 10⁹ l⁻¹), thrombocytopenia (PBC <100 × 10⁹ l⁻¹) or persistent anaemia (Hb <90 g l⁻¹ during the first 4 weeks post-transplant, based on previous experience that 90% of patients have Hb >90 g l⁻¹ by 1 month). All infectious episodes occurring in the first month after transplantation, and the results of microbiological investigation, were recorded. Cytomegalovirus (CMV) was diagnosed on the basis of compatible clinical illness and rising (fourfold) levels of pp65 antigen or a level >30. Bacterial infections were diagnosed by compatible clinical illness and positive microbiological culture. Median 2-h cyclosporin or trough tacrolimus concentrations were determined for each patient from the time of transplantation to the end of the first month. An average of 15 cyclosporin results (range 11–20) and 13 tacrolimus results (range 3–19) were collected per patient. For those patients with rejection during the first month, median concentrations of the associated calcineurin inhibitor were determined from the time of transplantation to the day of biopsy-proven rejection.

Statistical analysis

Normality of data distribution was assessed by the Kolmogorov–Smirnov test with Lilliefor's correction. Results were expressed as either mean ± SD for continuous parametric data, median (interquartile range, IQR) for continuous nonparametric data, or frequencies (percentages) for categorical data. Demographic, biochemical, pharmacokinetic and pharmacodynamic data for MMF +cyclosporin vs. MMF +tacrolimus treatment groups were compared by unpaired t-test, Mann–Whitney rank sum test or χ² test, depending on data type. Any significant covariates (e.g. age) were then included in subsequent analyses. The initial analysis multiple linear regression used SigmaStat (Version 1.0; SPSS Inc., North Sydney, Australia). The free fractions of MPA were related to free and total MPAG AUC, plasma urea, urate, albumin, globulin, bilirubin, glucose, and liver enzymes. Total protein was a multicolinear variable with albumin, so was omitted from the analysis. A probability value <0.05 was considered to be statistically significant. Any significant variables were further assessed for correlations with the free fraction of MPA using Spearman correlation analysis. These analyses were repeated for MPAG. Differences between the free MPA AUC₀₋₆ values of patients with and without haematological and infectious outcomes were determined by unpaired t-test. T-tests were also used to assess any statistically significant relation between free MPA AUC and donor type and kidney function. Linear regression analysis was used to compare the free fraction of MPA and MPAG calculated at each time point (FF_single) compared with that calculated from the full four time-point AUC (FF_full). Prediction errors were then calculated for each patient using the equation: (FF_single − FF_full)/FF_full · 100%. Mean prediction error (with 95% confidence intervals) was calculated as the arithmetic mean of the prediction errors for each patient [30]. Acute rejection rates were assessed by multivariate logistic regression, using rejection as the dependent variable and median drug concentration with either free, or total MPA AUC₀₋₆ as covariates.

Results

Baseline characteristics

A summary of baseline patient demographics and transplant characteristics is shown in Table 1. Patients in the cyclosporin cohort were significantly older and had a significantly higher median plasma bilirubin concentration than those in the tacrolimus group (median [IQR] = 13 [10–18] and 7 [5–12] µmol l⁻¹, respectively; P = 0.01). Otherwise there were no other significant differences between groups relating to race, sex, body weight, body mass index (BMI), diabetes, CAPD, transplant type, HLAM, PRA, ischaemic time, creatinine clearance, graft function or blood biochemistry.

Table 1.

Baseline demographics of the study population

	Total(n = 42)	Cyclosporin(n = 32)	Tacrolimus(n = 10)	P-value
Demographics
Age (years)	44.3 ± 13.1	46.9 ± 12.7	35.5 ± 10.2	0.02
Caucasian	41 (98%)	31 (97%)	10 (100%)	0.89
Male sex	24 (57%)	20 (63%)	4 (40%)	0.29
Body weight (kg)	72.9 ± 14.8	74.3 ± 15.2	68.6 ± 13.4	0.29
BMI (kg m−2)	25.3 ± 3.9	25.7 ± 3.9	24.1 ± 3.8	0.26
Diabetes	5 (12%)	4 (10%)	1 (10%)	0.92
CAPD pretransplant	14 (33%)	12 (38%)	2 (20%)	0.41
Cadaveric donor	29 (69%)	24 (75%)	5 (50%)	0.24
First transplant	37 (88%)	31 (97%)	6 (60%)	0.08
HLA mismatch	3 [2–5]	3 [2–5]	[2–4]	0.86
Peak PRA	0 [0–2]	0 [0–2]	0 [0–2]	0.94
Ischaemic time (h)	8.7 ± 5.9	9.5 ± 5.9	6.1 ± 5.2	0.11
Creatinine (mmol l−1)	0.16 [0.11–0.26]	0.16 [0.11–0.23]	0.14 [0.11–0.35]	0.85
CC (ml min−1)	40.1 ± 22.0	39.8 ± 19.4	41.0 ± 30.1	0.89
Delayed graft function	4 (10%)	3 (9%)	1 (10%)	0.99

Results are expressed as mean ±SD, median [IQR] or frequency (percentage), depending on data type. BMI, Body mass index; CAPD, continuous ambulatory peritoneal dialysis; HLA, haplotype mismatch; PRA, panel reactive bodies; CC, creatinine clearance. Delayed graft function is defined as the need for dialysis post-transplant.

MPA pharmacokinetics

There was considerable intersubject variability in free MPA AUC₀₋₆[coefficient of variation (CV%) = 67%] and total MPA AUC₀₋₆ (CV%= 42%) (Table 2). MPA free fraction varied (between patients) 11-fold, from 1.6% to 18.3%. The pharmacokinetic variation in total and free MPA was greater for patients on cyclosporin (12- to 18-fold variation) than on tacrolimus (four- to fivefold). There were no significant differences in mean total, or free MPA AUC₀₋₆, or median MPA free fraction between cyclosporin- and tacrolimus-treated groups (Table 2).

Table 2.

Pharmacokinetic measures of free, and total mycophenolic acid in adult renal allograft recipients on day 5 after transplantation

	Total(n = 42)	Cyclosporin(n = 32)	Tacrolimus(n = 10)	P-value[95% CI]
Total mycophenolic acid
AUC0−6 (mg h−1 l−1)	22.0 ± 9.2(3.7–49.1)	20.8 ± 8.4(3.7–44.4)	25.6 ± 11.1(10.7–49.1)	0.16[− 11.4, 1.89]
C0 (mg l−1)	Median 0.9[0.6–1.6]	Median 0.8[0.6–1.3]	Median 1.6[1.4–3.0]	0.002
Free mycophenolic acid
AUC0−6 (mg h−1 l−1)	1.2 ± 0.8(0.2–2.0)	1.2 ± 0.8(0.2–3.7)	1.1 ± 0.6(0.5–2.0)	0.82[− 0.5, 0.6]
C0 (µg l−1)	Median 43.1[18.9–75.0]	Median 37.1[12.9–73.8]	Median 61.5[43.8–76.0]	0.10
Free mycophenolic acid fraction*
% free/total	Median 4.8[3.0–8.1]	Median 5.0[3.1–8.5]	Median 4.2[2.9–6.9]	0.49

Data are mean ± SD and range of values, or median [IQR], depending on data type. 95% CI is the confidence interval on the difference between means. AUC₀₋₆, area under the curve from 0 to 6 h; C₀, trough concentration.

^*Value calculated from AUC₀₋₆ results only.

The 6-h pharmacokinetic profile of mean total and free MPA concentrations from all patients is shown in Figure 1. Maximum total, and free MPA concentrations (4.9 ± 3.7 mg l⁻¹ and 0.29 ± 0.27 mg l⁻¹, respectively) occurred at 1 h after dose using this four time-point sampling profile. There was no correlation of MPA trough concentration with MPA AUC₀₋₆ for total or free MPA (r² = 0.16 and 0.11, respectively). Total and free MPA AUC₀₋₆ did not differ between patients receiving cadaveric vs. live donor organs (P = 0.13 and 0.67, respectively).

Figure 1.

Six-hour pharmacokinetic profiles for free (▴) and total (▾) mycophenolic acid with free (•) and total (▪) mycophenolic acid glucuronide. Data shown are mean ± SD

MPAG pharmacokinetics

There were no significant differences between the cyclosporin- and tacrolimus-treated groups with regard to total or free MPAG AUC₀₋₆, or the MPAG free fraction (Table 3). Intersubject variability of total and free MPAG AUC₀₋₆ was, as for MPA, higher in the cyclosporin-treated patient group (eight- to 14-fold) compared with those in the tacrolimus-treated group (seven- to 10-fold). MPAG free fraction ranged from 17.4% to 54.1%. The 6-h pharmacokinetic profile of total and free MPAG concentrations is shown in Figure 1. There was a positive correlation of MPAG trough concentration with the MPAG AUC₀₋₆ for the wide range of concentrations measured (total, 18–403 mg l⁻¹, r² = 0.95; free, 3–145 mg l⁻¹, r² = 0.98).

Table 3.

Pharmacokinetic measures of free, and total MPAG in adult renal transplant recipients on day 5 after transplantation

	Total (n = 42)	Cyclosporin (n = 32)	Tacrolimus (n = 10)	P-value
Total mycophenolic acid glucuronide
AUC0−6 (mg h−1 l−1)	Median 714.3 [462.8–1001.9]	Median 714.3 [474.6–1048.3]	Median 694.1 [462.8–824.7]	0.54
C0 (mg l−1)	Median 87.3 [51.3–143.4]	Median 87.8 [54.5–164.8]	Median 70.2 [47.8–143.4]	0.59
Free mycophenolic acid glucuronide
AUC0−6 (mg h−1 l−1)	Median 177.2 [121.5–328.2]	Median 177.2 [131.0–339.0]	Median 167.2 [109.0–214.5]	0.65
C0 (mg l−1)	Median 18.4 [11.4–43.0]	Median 18.4 [11.3–45.7]	Median 17.8 [11.5–26.3]	0.76
Free mycophenolic acid glucuronide fraction*
% free/total	Median 27.4 [23.3–33.1]	Median 28.2 [23.2–32.9]	Median 26.9 [23.6–35.8]	0.63

Data are expressed as median [IQR]. AUC₀₋₆, Area under the curve from 0 to 6 h; C₀, trough concentration.

^*Value calculated from AUC₀₋₆ results only.

Ratios of total MPA to total MPAG were significantly higher in tacrolimus-treated patients than in those on cyclosporin therapy (medians 0.028 and 0.008, respectively, P < 0.01; Mann–Whitney rank sum test).

Protein binding

A summary of the parameters that significantly influenced the protein binding of MPA and its glucuronide is shown in Table 4. The free fractions of MPA and its glucuronide were both positively correlated with urea and creatinine concentrations (r > 0.41) and negatively correlated with albumin concentration. MPA free fraction was highly influenced by free and total MPAG AUC₀₋₆ (r > 0.51).

Table 4.

Spearman correlation coefficients and P-values showing the parameters that were significantly related to the free fractions of mycophenolic acid or mycophenolic acid glucuronide in 42 patients

	Free fraction
	Mycophenolic acid		Mycophenolic acid glucuronide
Parameter	Correlation	P-value	Correlation	P-value
Creatinine	0.46	0.002	0.41	0.008
Urea	0.53	<0.0001	0.41	0.007
Albumin	−0.54	<0.0001	−0.31	0.049
Total mycophenolic acid glucuronide AUC0−6	0.51	<0.0001	–	–
Free mycophenolic acid glucuronide AUC0−6	0.60	<0.0001	–	–

Results were calculated using Spearman correlation tests. AUC₀₋₆, Area under the curve from 0 to 6 h.

We previously found plasma albumin concentrations = 31 g l⁻¹ to be associated with elevations in the free fraction of MPA [31]. The mean free fraction for patients in the present study with albumin ≤31 g l⁻¹[mean (albumin) = 27.5 ± 3.4 g l⁻¹, n = 26] was 70% higher than in patients with albumin >31 g l⁻¹[mean (albumin) = 34.0 ± 2.3 g l⁻¹, n = 16][free fraction = 6.9 ± 3.6% and 4.1 ± 3.3%, respectively; P = 0.001; 95% confidence interval (CI) for the difference 1.9, 4.2]. The protein binding data were further analysed with patients allocated to one of four cohorts according to their primary immunosuppression and the predetermined cut-off value for albumin, i.e. cyclosporin- or tacrolimus-treated with albumin concentrations ≤31 g l⁻¹, vs. >31 g l⁻¹. Two-way anova revealed that patients with plasma albumin concentrations <31 g l⁻¹ had higher trough MPA free fractions if on cyclosporin compared with those on tacrolimus (two-way anova, P = 0.03). These results are consistent with the lower median trough concentrations of total MPA in the cyclosporin- vs. tacrolimus-treated transplant recipients.

Rejection

Only seven patients experienced acute rejection with the median (IQR) time to rejection 14 (7.5–19.5) days after transplant. The free MPA AUC₀₋₆ of rejectors was not significantly different from that of nonrejectors (rejectors median 1.4 mg h⁻¹ l⁻¹, IQR 1.3–2.1 mg h⁻¹ l⁻¹; nonrejectors median 0.9 mg h⁻¹ l⁻¹, respectively, IQR 0.6–1.5 mg h⁻¹ l⁻¹; P = 0.07). Logistic regression of these data was also performed separately on the data from each of the cyclosporin and tacrolimus cohorts, with median calcineurin inhibitor concentration included as a continuous variable within each model. This again showed no difference between rejectors and nonrejectors (P = 0.26). The total MPA AUC₀₋₆ of rejectors was not significantly different from that of nonrejectors (rejectors 18.2 ± 7.9 mg h⁻¹ l⁻¹; nonrejectors 22.7 ± 9.4 mg h⁻¹ l⁻¹;P = 0.25). Neither cyclosporin nor tacrolimus concentrations were different between rejectors and nonrejectors (P = 0.53 and P = 0.68, respectively).

Toxicity

Fourteen patients experienced adverse effects. Of these, four experienced gastrointestinal events, nine had haematological side-effects (some with infection) and one patient experienced both blood-associated and gastrointestinal events. Of infections encountered, 33% were CMV, 17% Methicillin-resistant Staphylococcus aureus bacteraemia, 17% urinary tract infection and 33% wound infection/cellulitis. There was no significant difference in free MPA AUC₀₋₆ between patients with and without gastrointestinal events (P = 0.65), nor with and without anaemia (P = 0.43). Those patients who experienced one or more haematological events including thrombocytopenia and leukopenia or infectious events had significantly higher free MPA AUC₀₋₆ (but not total MPA AUC₀₋₆ values; P = 0.18) compared with patients without such side-effects (mean ± SD 1.9 ± 0.3 mg h⁻¹ l⁻¹ and 1.1 ± 0.1 mg h⁻¹ l⁻¹, respectively, P = 0.0043; 95% CI for the difference 0.3, 1.4) (Figure 2).

Figure 2.

Free mycophenolic acid AUC₀₋₆ in patients who experienced leukopenia, thrombocytopenia or infection (present) vs. those patients who experienced none of these adverse events (absent); *t-test

Discussion

This study has defined the pharmacokinetics and pharmacodynamics of both total and free (unbound) concentrations of MPA and its major glucuronide metabolite (MPAG) in a cohort of renal transplant recipients early post-transplant. The pharmacokinetic variability of MPA and its glucuronide using 6-h AUC was comparable to that using 12-h AUC [6–8, 14]. Contributory factors may include concomitant medication effects as well as renal and hepatic disease (e.g. hypoalbuminaemia) [7, 9]. In the present study a much wider range of free and total MPA AUC values for patients treated with cyclosporin compared with tacrolimus was apparent [32]. This difference in variability that was also apparent for MPAG has not been previously reported.

Patients in the cyclosporin-treated group had mean total MPA trough concentrations (for the same dose) approximately half that of patients in the tacrolimus-treated group. Reduced MPA concentrations were probably due to reduced MPA enterohepatic recycling associated with cyclosporin [12, 33]. Variation in enterohepatic recycling could also account for the considerably increased variability in AUC for the cyclosporin-treated group. The higher ratio of total MPA to total MPAG in tacrolimus-treated patients compared with those on cyclosporin therapy that was recorded in the current, and other studies [6, 34, 35] is in keeping with this mechanism.

The average free MPA AUC₀₋₁₂ has been reported to range between 0.8 and 1.8 mg h⁻¹ l⁻¹ for adult renal recipients on up to 3 g day⁻¹ oral MMF plus cyclosporin [7, 9], which is in line with the values obtained in the present study for AUC₀₋₆.

MPA free fraction in this study (5.8 ± 3.7%) is comparable to that reported previously on 7 days post-transplant (7.7 ± 2.1%, n = 8) [7], and in patients with chronic renal insufficiency (5.8 ± 2.7%, n = 8) [20]. Stable renal allograft patients have MPA free fractions of 1.2–3.1%[36]. One patient in the current study had a MPA free fraction of 18.3%, the highest reported to date. MPA free fractions >10% are extremely uncommon, with only one other published case (13.3%) [37]. Almost all the patients (37/42) in the present study had MPA free fraction values >2.5%, the reported upper limit for stable renal patients [38]. There was no significant difference between the MPA free fractions of cyclosporin- vs. tacrolimus-treated groups. Our study supports in vitro data that MPA binding is not altered by these common immunosuppressant medications [19]. The lack of difference between recipients of cadaveric vs. live donor organs suggests that the groups were well matched with respect to degree of renal dysfunction.

Accumulation of MPAG in renal impairment could account for the variation in its pharmacokinetics. The total MPAG concentrations measured in this study were similar to those reported by Johnson et al. [22] (100–180 mg l⁻¹) in the first month post-transplant for patients on equivalent doses of MMF (Table 3). The average MPAG AUC₀₋₁₂ has been reported to range between 2.1 and 5.7 mg h⁻¹ l⁻¹ for adult renal recipients in the first week post-transplant on up to 3 g day⁻¹ oral MMF plus cyclosporin [7, 9]. The MPAG AUC₀₋₆ values measured in this study were lower (as expected for the 6-h collection period), around 0.7 mg h⁻¹ l⁻¹, with no difference between cyclosporin- and tacrolimus-treated groups.

Predose MPAG total concentrations of 36–199 mg l⁻¹ and free concentrations of 8–58 mg l⁻¹ in stable renal allograft recipients treated with MMF and cyclosporin have recently been published [36]. These values were lower than those measured in the current study (17–403 mg l⁻¹ and 3–145 mg l⁻¹, respectively) which included patients with a range of renal function (stable to severely impaired). The range of MPAG free fraction values measured in the current study (17–54%) is higher than previously reported in healthy and/or stable renal transplant recipients [9, 36].

Another novel finding of the current study was the strong relation between MPAG trough concentration with AUC₀₋₆ for both free (r² = 0.98) and total glucuronide concentrations (r² = 0.95). Whilst this has not been shown for the phenolic glucuronide, a similar relationship has been reported for the acyl glucuronide metabolite (r² = 0.93) [39]. This suggests that predose samples can therefore be used as surrogates for the AUC of the metabolites.

The absence of a relationship between free or total MPA AUC, as measured on day 5 after transplant, or tacrolimus or cyclosporin median concentrations and rejection within the first month in the present study is probably due to insufficient numbers of rejection episodes. The time elapsed between MPA measurement and clinical event also makes this type of assessment difficult; however, this is the clinical reality often confronting prescribers and it is worth investigating whether early measures of immunosuppressant exposure can predict such outcomes. A pharmacokinetic/dynamic relationship has been described previously between total MPA AUC and the risk of rejection in a study on a small number of adult and renal transplant recipients receiving MMF in combination with cyclosporin [40].

Two studies have found that renal dysfunction is associated with higher free fraction and higher AUC values of MPA [7, 9]. Another study found that the percentage of free MPA, but not total, correlated with red blood cell and leucocyte count [41]. Of interest in the current study is the significantly higher free but not total MPA AUC₀₋₆ value found in patients with haematological or infectious events. These data provide further support for measuring the active, unbound drug in preference to total concentrations. It also demonstrates the potential value of using measures of early exposure as predictors of later clinical events. An increased risk of leukopenia and infection has been previously reported with high free, but not total, MPA AUC₀₋₁₂ values [7, 9]. Oellerich et al. [10] observed this greater risk in patients with free MPA AUC₀₋₁₂ values = 0.6 mg h⁻¹ l⁻¹.

The absence of a relationship between free MPA AUC₀₋₆ and the incidence of gastrointestinal effects is consistent with previous findings in paediatric and adult renal transplant recipients [10, 13, 21] and is in keeping with the hypothesis that gastrointestinal adverse events may be related to local rather than systemic concentrations.

It has been previously shown that renal function, albumin concentration and MPAG all may influence protein binding of MPA [7–9, 20]. As with other highly protein-bound drugs [42, 43], this study confirmed that MPA free fraction was inversely correlated with albumin. Furthermore, MPAG, which displaces MPA from albumin, and renal function (creatinine) were positively correlated with free fraction. The fact that albumin had a greater influence on MPA than on MPAG binding may reflect their relative binding affinities to this ligand.

In conclusion, this study of post-transplant recipients on MMF with a wide range of renal function has highlighted the large interindividual variation in both free and total MPA pharmacokinetics early after transplant. This variability was greater in cyclosporin-treated than tacrolimus-treated subjects and was similar with free and total MPAG. We have found that the majority of patients had significantly altered MPA and MPAG protein binding. The major influences on protein binding were albumin concentration and renal impairment. There are inherent limitations of small patient numbers and documented events; however, the findings suggest that early measurement of MPA exposure (particularly of free drug) is a better predictor of haematological or infectious adverse events than total concentrations. Further investigations of the complex interrelations between plasma biochemistry, MPA pharmacokinetics and pharmacodynamic outcomes are warranted, especially with regard to hypoalbuminaemic patients.