 |
||||
| 2007 JASN IMPACT FACTOR 7.111 | HOME AUTHOR INFO EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP | |||
| CURRENT ISSUE | ARCHIVES | JASN Express | ONLINE SUBMISSION | |
|
||||
| 2007 JASN IMPACT FACTOR 7.111 | HOME AUTHOR INFO EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP | |||
| CURRENT ISSUE | ARCHIVES | JASN Express | ONLINE SUBMISSION | |
*Division of Pediatric Nephrology, University Children’s Hospital, Heidelberg, Germany; Department of Clinical Chemistry, Georg-August University, Göttingen, Germany; and University Children’s Hospital Freiburg, Germany.
Correspondence to: Dr. Burkhard Tönshoff, Division of Pediatric Nephrology, University Children’s Hospital, Im Neuenheimer Feld 150, 69120 Heidelberg, Germany. Phone: 49-6221-562311; Fax: 49-6221-564203; E-mail: Burkhard_Toenshoff{at}med.uni-heidelberg.de
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MMF is rapidly metabolized in vivo to its active constituent, mycophenolic acid (MPA), a reversible, uncompetitive inhibitor of inosine monophosphate dehydrogenase (IMPDH). Inhibition of IMPDH-II in activated lymphocytes causes a reduction in intracellular guanine nucleotide pools and leads to an arrest of lymphocyte proliferation. MPA is extensively bound to albumin, with a range of protein binding of 97% to 99% in patients with normal renal and liver function (3–5). On the basis of in vitro investigations, free MPA concentrations are presumed to be responsible for its immunosuppressive action (4). The primary metabolite of MPA is the phenolic glucuronide 7–0–MPAG. Two further metabolites have been identified in humans, namely the acyl glucuronide (AcMPAG) and the phenolic glucoside of MPA (6). Of these three metabolites, only the acyl glucuronide is capable of inhibiting human IMPDH-II in vitro (7).
The pharmacokinetics of MPA shows large interindividual and intraindividual variability. Pharmacokinetic monitoring of MPA with the aim of optimizing the dosage of this drug to achieve adequate immunosuppression with minimized risk of graft rejection or toxicity has therefore been under investigation (8). In adult renal transplant recipients on cyclosporine, MMF, and steroids, pharmacokinetic/pharmacodynamic (PK/PD) relationships between MPA-AUC or predose MPA levels and the risk of acute rejection have been established on the basis of measurements by reverse-phase HPLC with ultraviolet detection (9,10).
The interpretation of the PK/PD relationship of MPA in different patient populations can be influenced by several factors, such as the nature of the organ transplanted, the age of the patient, use of concomitant immunosuppressive therapy, protein binding, the presence of active metabolites, and the assay used. The purposes of this investigation were therefore (1) to establish the PK/PD relationships for total and free MPA in pediatric renal transplant recipients on an immunosuppressive regimen with cyclosporin A, MMF, and corticosteroids and (2) to establish the therapeutic ranges for MPA-AUC0–12 values and predose MPA levels in this patient population.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Pharmacokinetic Protocol
Patients were studied after informed (parental) consent was obtained. Blood samples for pharmacokinetic assessments were drawn on days 7 and 21 posttransplant (initial phase) and 3 and 6 mo posttransplant (stable phase). It was mandatory that all patients had at least 2 full days of the same MMF dose given twice a day before pharmacokinetic investigations. The study was performed under in-patient conditions, starting in the morning. Patients were required to fast from 10 p.m. the night before sampling until after the 75-min sample had been obtained on the following morning. Blood samples were collected at the following times: before dosing and 20 min, 40 min, 75 min, and 2, 4, 6, 8, and 12 h after dosing. The study protocol was approved by the local ethics committee of each contributing center. All blood samples were collected in tubes containing ethylenediaminetetraacetic acid as an anticoagulant. For determination of MPA, free MPA, and MPAG concentrations, plasma was separated and stored at -20°C until analysis.
Measurement of Total and Free MPA and MPAG
The procedure for the determination of total MPA and MPAG in plasma, as well as for the determination of the free concentration of MPA in plasma water has been described in detail elsewhere (5,15). The limit of quantification for free MPA (at 215 nm) was set at 10 µg/L because the imprecision at the detection limit of 5 µg/L was >20%. The within-day imprecision ranged from 6.5% (209.2 µg/L) to 11.6% (18.4 µg/L) and the between-days imprecision ranged from 7.2% (200.6 µg/L) to 14.6% (22.8 µg/L) for measurement of free MPA at 215 nm (19).
Pharmacokinetic Analyses
The following pharmacokinetic data for MPA, free MPA, and MPAG were determined: predose concentration (C0), time to maximum concentration (tmax [h]), maximum concentration (Cmax [mg/L]), area under the curve (AUC) from 0 to 12 h (mg x h/L) using the linear trapezoidal rule, and the evening predose (i.e., the 12-h) concentration (C12 [mg/L]). An abbreviated three-point AUC based on the sampling times 0 min, 75 min, and 4 h (MPA-AUC0,75 min,4 h) was calculated according to a previously published algorithm (20): estimated AUC = 11.8 + 3.71 x C0 + 1.33 x C75 min + 3.9 x C4 h, where C0 is the predose MPA concentration, and C75 min and C4 h are MPA concentrations 75 min and 4 h, respectively, after MMF dosing. In addition, a published algorithm based on an empiric equation using a limited sampling strategy up to 2 h postdose (MPA-AUC0–2) was taken for the calculation of the full AUC (21). The pharmacokinetic analysis was performed using the computer program BiAS (Epsilon-Verlag Hochheim-Darmstadt, Germany).
Acute Rejection Episodes
The primary outcome criterion for the determination of the PK/PD relationship for MPA was the occurrence of acute rejection episodes over the 6-mo study period after transplantation. Fifteen of 54 patients experienced at least one acute rejection episode during the 6-mo study period; two of these patients had two rejection episodes, resulting in a total number of 17 acute rejection episodes. Twelve of 17 acute rejection episodes were biopsy-proven, and histologic examination and classification of a core biopsy was done according to the Banff criteria (16). If a biopsy was logistically impossible or clinically contraindicated, the diagnosis of "presumed rejection" was based on clinical judgment (supported by one or more of the following clinical findings: increased body temperature, graft swelling, graft tenderness, rise in serum creatinine level of more than 20% from the baseline level, or oliguria). The clinical diagnoses were made without any knowledge of the MPA concentration data. Acute rejection episodes were treated initially with high-dose intravenous corticosteroids (methylprednisolone, 400, 200, 200, and 100 mg/m2 BSA on 4 consecutive days, respectively, or 300 mg/m2 BSA for 3 to 5 d). If the rejection episode failed to respond to this therapy, treatment with OKT3 or antithymocyte globulin was started or patients were converted from cyclosporine to tacrolimus according to center practice. The acute rejection episodes occurred 27 d (range, 8 to 170 d) after renal transplantation; 14 of 17 rejections occurred within the first 70 d posttransplant.
Adverse Event Monitoring
Adverse events, defined as an abnormal change in physical signs, symptoms, or laboratory values whether or not deemed to be causally related with the study medication, were recorded throughout the study when reported by a patient or noted by an investigator. Thrombocytopenia was defined as a thrombocyte count <150 x 1012/L. Leukopenia was defined as a granulocyte count <2000/µl and graded according to its severity: mild (1600 to 1999 granulocytes); moderate (1000 to 1599 granulocytes) (n = 5); severe (500 to 999 granulocytes). Diarrhea was graded as follows: mild diarrhea (n = 7), transient diarrhea 
2 d; moderate diarrhea (n = 3), tolerable diarrhea lasting >2 d; and severe diarrhea (n = 1), intolerable diarrhea requiring therapy. Infections were classified as moderate when requiring specific antibiotic or antiviral therapy and as severe when requiring hospitalization. The following infections were recorded: herpes labialis (n = 5), herpes zoster (n = 1), oral mucocutaneous candidiasis (n = 1), urinary tract infection (n = 8), cytomegalovirus (CMV)–pneumonia (n = 1), CMV-colitis (n = 1), bacterial septicemia (n = 3), pneumonia (n = 1), pharyngitis (n = 1), febrile viral infection (n = 4).
Statistical Analyses
The Shapiro-Wilks test was used to confirm normal distribution of data (17). Because some parameters were normally distributed and some were not, data in Table 1 and 2 are given as median (range). For comparison of more than two groups, one-way ANOVA on repeated measurements followed by all pairwise comparison (Student Newman-Keuls test) was used. Forward stepwise regression analysis was performed to assess which variables contribute independently to the prediction of MPA-AUC0–12 values. Differences were considered to be statistically significant at a P < 0.05.
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Because a full MPA-AUC requires at least 8 blood samples during a 12-h dose interval and is therefore impractical in clinical routine practice, we investigated whether a single time point MPA concentration or an abbreviated AUC derived from a limited number of samples correlate with the full MPA-AUC. Only a moderate correlation was observed between either the predose MPA trough concentration (C0) or the 12-h evening trough concentration (C12) and the full MPA-AUC0–12. Although the C0 concentration may have been taken after a more variable time interval subsequent to the previous dose than the C12 concentration that was sampled exactly 12 h after a supervised dose, the correlation of C12 with the full AUC was not superior to that of C0 (Table 4). There was also only a moderate correlation between the MPA peak concentration (Cmax) and the respective MPA-AUC0–12 (Table 4). Because MPA Cmax increases during the first 3 mo posttransplant more markedly than the MPA predose concentrations (Table 2), the coefficients of correlation of MPA Cmax with the respective full AUC in the stable phase posttransplant were higher and the coefficients of correlation of MPA predose levels were lower than in the initial phase posttransplant. The abbreviated MPA-AUC0,75 min,4 h gave a reasonable correlation with the full AUC that was somewhat superior to that seen between the abbreviated profile MPA-AUC0–2 and the full AUC (Table 4).
|
|
To establish which pharmacokinetic parameter is the best predictor for the risk of acute rejection, ROC curves were computed for each of the four PK parameters. For these calculations, the mean values from the 1-wk and 3-wk sampling time points were taken for each variable and patient. Diagnostic sensitivities (true positive) were calculated for each individual PK value as the fraction of those patients with an acute rejection who had a value below this discrimination threshold. The corresponding diagnostic specificities (false negatives) were calculated as the fraction of patients with no rejection episode who had a value below this decision value. The ROC plots of sensitivity versus 1-specificity for the PK variables AUC0–12 and C12 are shown in Figure 2, and the statistical comparison of the areas under the ROC curves is given in Table 5. C12, AUC0–12, and the two abbreviated AUC estimations were able to discriminate patients with acute rejections from patients with no rejection. An AUC0–12 of 33.8 mg x h/L had a diagnostic sensitivity of 75.0% and a diagnostic specificity of 64.3%. For the PK parameter C12, a concentration of 1.2 mg/L gave a sensitivity of 83.3% and a specificity of 64.3%. In contrast, the PK variables derived from the measurements of free MPA were not effective for the discrimination of rejectors from nonrejectors (Table 5).
|
|
|
|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The incidence of the MMF-related adverse events, diarrhea, vomiting, or abdominal pain, was not associated with MPA PK variables, which is consistent with previous findings in adult renal transplant recipients (10). These adverse events might also reflect local effects of MMF in the gastrointestinal system and not only systemic exposure. The adverse events, leukopenia and/or infections, were also not associated with any of the PK variables derived from total MPA measurements but with the AUC0–12 of free MPA. Most of these adverse events were associated with free MPA-AUC0–12 levels >0.4 mg x h/L (Figure 3). MPA is extensively bound to albumin, with a range of protein binding of 97% to 99% in patients with normal renal and liver function (3–5). Factors that alter protein binding can affect the relationship between total concentration and free concentration. The pharmacologically inactive phenolic glucuronide metabolite of MPA, MPAG, competes for the albumin binding sites and can increase the fraction of unbound active MPA when present at high concentrations, for example as a consequence of renal failure (5,19). The presence of hyperbilirubinemia, uremia, low pH, and low albumin concentrations also can decrease MPA binding (3–5,19). For this reason, under circumstances of perturbed binding, the interpretation of total MPA plasma concentrations must take into account the altered MPA binding. This observation of an increased toxic risk of MMF therapy in patients with hypoalbuminemia and/or reduced renal function is clinically important, because MMF is increasingly used for immunosuppressive therapy in patients with chronic glomerulonephritis and the nephrotic syndrome (22). In this context, it is of interest to note that Kaplan et al. (23) reported a case of severe leukopenia that was associated with renal insufficiency in a pancreas/renal transplant recipient receiving 0.75 g of MMF twice daily. In this case, the free MPA-AUC0–12 was substantially elevated (5.1 mg x h/L) and the total MPA-AUC0–12 was in the range observed for stable renal transplant recipients.
The apparent discrepancy that the AUC0–12 of total MPA was associated with the incidence of acute rejection episodes, whereas the AUC0–12 of free MPA was associated with side effects reflecting the systemic exposure to MMF deserves comment. It is possible that a low protein binding of MPA favors its access from the circulation to the bone marrow, where it induces leukopenia by inhibition of leukocyte maturation from precursor cells, whereas inhibition of proliferation of circulating lymphocytes is mainly responsible for its immunosuppressive effect. A recent preliminary report in stable kidney transplant recipients has confirmed the relation between free MPA levels and the incidence of leukopenia (24).
In agreement with our previous reports in smaller cohorts of patients (5,11), there was a large interindividual variation of PK data (Table 2). In the first week posttransplant, but not at later PK sampling periods, low MPA-AUC0–12 values were associated with young age, low serum albumin, and decreased transplant function. This transient association between impaired renal function and low MPA-AUC0–12 values in the initial but not stable phase posttransplant has recently been reported in adult renal transplant recipients (25). Patients with decreased renal function have an increased MPA free fraction because the accumulation of the renal metabolite MPAG competes with MPA for binding sites on albumin (5,26,27). Hence, the most likely explanation for the association between low MPA-AUC0–12 values and impaired renal function is that decreased protein binding in these patients leads to an enhanced MPA metabolism, because it is the free drug that is primarily available for metabolism and excretion. Other possible causes of the interindividual variability of MPA-AUC0–12 values include differences in the absorption of MMF and/or in the metabolism via the glucuronidation pathway.
Our study has demonstrated a significant relationship between both MPA-AUC0–12 and predose MPA levels and the risk of acute rejection. This observation provides some encouragement that therapeutic drug monitoring with individualized dosing of MMF could optimize the usage of this immunosuppressive drug. There are additional lines of evidence that favor such a monitoring. First, MPA displays highly variable interindividual pharmacokinetics; in the initial phase posttransplant, there is also considerable intraindividual PK variability (3,5,8,11). Second, there is a time-dependent increase in MPA-AUC0–12 in both adult and pediatric renal transplant recipients on concomitant immunosuppression with cyclosporin A and corticosteroids (8,11) that might require time-dependent dose adjustments of MMF. Third, the pharmacokinetics of MPA is also dependent on the concomitant immunosuppression (28). However, it has to be emphasized that the value of therapeutic drug monitoring of MPA has not been formally tested yet. Only a prospective trial in which kidney transplant recipients treated with the same triple regimen are randomized for treatment with a fixed dose of MMF versus treatment with an MMF dose adjusted to the results of plasma concentration monitoring could establish the scientific basis of therapeutic drug monitoring of MPA. Such rigorous criteria are required to justify the additional effort and expense involved, before a general therapeutic drug monitoring of MPA can be recommended. Moreover, attempts to define the PK/PD relationship for gastrointestinal disturbances have not been successful, which is consistent with the results in this study. Instead, the dose of MMF provides the most predictive information for adverse events such as diarrhea (10), presumably because of the local toxicity of MPA on the gastrointestinal epithelium. However, given the association between MPA PK variables and the risk of acute rejection, therapeutic drug monitoring of MPA in the initial phase posttransplant could be helpful in selected high-risk patients. It would be reasonable to aim at achieving MPA-AUC values or predose MPA levels somewhat higher than those achieved in the lower percentiles by individual adjustment of the MMF dose. For such an approach, the patient-specific and immunosuppressive regimen–specific percentiles for MPA-AUC and MPA predose levels from this study could serve as a guideline for optimization of MMF therapy in pediatric renal transplant recipients. On the basis of data from this and other studies, a consensus is arising that for minimizing the risk of rejection after transplantation, total MPA-AUC values in the early stages posttransplant should be maintained within the therapeutic window of 30 to 60 mg x h/L, based on HPLC data, or the predose concentration in the range of 1 to 3.5 mg/L. Because some fluctuations are to be expected as a consequence of the enterohepatic circulation of MPA, extreme values or a large change should be verified by repeat measurement before a change in the MMF dose is undertaken.
The question arises of which PK parameter of total and free MPA is best suited for therapeutic drug monitoring of MMF therapy in clinical practice. In this study, the parameters MPA-AUC0–12, MPA-AUC0,75 min,4 h, and MPA C12 were comparable for assessing the risk of acute rejection episodes, whereas MPA C0 and MPA-AUC0–2 had a somewhat poorer predictive value (Table 5). In two previous studies in adult renal transplant recipients, the MPA-AUC was a better predictor of outcome than the predose concentration (21,29). Predose levels are more convenient than measurements of AUC that are complex and increase the cost of routine clinical monitoring. However, the predose C12, i.e., the concentration measured 12 h after a supervised dose, is not the predose sample commonly collected in a clinic. This is the C0 predose concentration, which may be sampled at a more variable time interval after the last dose. Hence, the calculation of AUC using a three-point limited sampling strategy up to 4 h after MMF dosing (MPA-AUC0,75 min,4 h) appears to be a more precise approach for the assessment of the risk of acute rejection than the predose C0 level.
In conclusion, we have observed an association between the risk of acute rejection episodes and MPA-AUC values or MPA predose levels in pediatric renal transplant recipients on an immunosuppressive triple drug therapy with cyclosporin A, MMF, and corticosteroids. We have established 5th to 95th percentiles for these PK parameters on the basis of HPLC measurements in this specific patient population. These reference data could facilitate the therapeutic drug monitoring of MPA for optimization of MMF efficacy by steering patients away from the extreme values of MPA PK variables indicated by the lower percentiles. For the assessment of the toxic risk of MMF regarding leukopenia and/or infections, measurement of free MPA appears to be more appropriate.
|
Acknowledgments |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R. Hanvesakul, C. Kubal, S. Jham, E. Sarkar, K. Eardley, D. Adu, and P. Cockwell Increased incidence of infections following the late introduction of mycophenolate mofetil in renal transplant recipients Nephrol. Dial. Transplant., December 1, 2008; 23(12): 4049 - 4053. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Jacobson, J. Huang, N. Rydholm, M. Tran, T. DeFor, J. Tolar, and P. J. Orchard Higher Mycophenolate Dose Requirements in Children Undergoing Hematopoietic Cell Transplant (HCT) J. Clin. Pharmacol., April 1, 2008; 48(4): 485 - 494. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W.-P. Yau, A. Vathsala, H.-X. Lou, and E. Chan Is a standard fixed dose of mycophenolate mofetil ideal for all patients? Nephrol. Dial. Transplant., December 1, 2007; 22(12): 3638 - 3645. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. M. Shaw, M. Figurski, M. C. Milone, J. Trofe, and R. D. Bloom Therapeutic Drug Monitoring of Mycophenolic Acid Clin. J. Am. Soc. Nephrol., September 1, 2007; 2(5): 1062 - 1072. [Full Text] [PDF]
|
|
|
|
|
|
H. Jeong and B. Kaplan Therapeutic Monitoring of Mycophenolate Mofetil Clin. J. Am. Soc. Nephrol., January 1, 2007; 2(1): 184 - 191. [Full Text] [PDF]
|
|
|
|
 |
|
 O. Bernard, J. Tojcic, K. Journault, L. Perusse, and C. Guillemette Influence of Nonsynonymous Polymorphisms of UGT1A8 and UGT2B7 Metabolizing Enzymes on the Formation of Phenolic and Acyl Glucuronides of Mycophenolic Acid Drug Metab. Dispos., September 1, 2006; 34(9): 1539 - 1545. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. M. van Hest, R. A.A. Mathot, M. D. Pescovitz, R. Gordon, R. D. Mamelok, and T. van Gelder Explaining Variability in Mycophenolic Acid Exposure to Optimize Mycophenolate Mofetil Dosing: A Population Pharmacokinetic Meta-Analysis of Mycophenolic Acid in Renal Transplant Recipients J. Am. Soc. Nephrol., March 1, 2006; 17(3): 871 - 880. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Giaccone, J. S. McCune, M. B. Maris, T. A. Gooley, B. M. Sandmaier, J. T. Slattery, S. Cole, R. A. Nash, R. F. Storb, and G. E. Georges Pharmacodynamics of mycophenolate mofetil after nonmyeloablative conditioning and unrelated donor hematopoietic cell transplantation Blood, December 15, 2005; 106(13): 4381 - 4388. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. A. Jacobson, K. G. Green, and B. J. Hering Mycophenolate Mofetil in Islet Cell Transplant: Variable Pharmacokinetics but Good Correlation Between Total and Unbound Concentrations J. Clin. Pharmacol., August 1, 2005; 45(8): 901 - 909. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Hazzan, M. Labalette, M. C. Copin, F. Glowacki, F. Provot, F.-R. Pruv, and C. Noel Predictive Factors of Acute Rejection after Early Cyclosporine Withdrawal in Renal Transplant Recipients Who Receive Mycophenolate Mofetil: Results from a Prospective, Randomized Trial J. Am. Soc. Nephrol., August 1, 2005; 16(8): 2509 - 2516. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. M. Annesley and L. T. Clayton Quantification of Mycophenolic Acid and Glucuronide Metabolite in Human Serum by HPLC-Tandem Mass Spectrometry Clin. Chem., May 1, 2005; 51(5): 872 - 877. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. G Morris Immunosuppressant Drug Monitoring: Is the Laboratory Meeting Clinical Expectations? Ann. Pharmacother., January 1, 2005; 39(1): 119 - 127. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Streit, M. Shipkova, V. W. Armstrong, and M. Oellerich Validation of a Rapid and Sensitive Liquid Chromatography-Tandem Mass Spectrometry Method for Free and Total Mycophenolic Acid Clin. Chem., January 1, 2004; 50(1): 152 - 159. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Neumann, M. Haidinger, H. Jager, H. Grutzmacher, A. Griesmacher, M. M. Muller, P. M. Bayer, and F. Thomas Meisl Pharmacokinetics of Mycophenolate Mofetil in Patients with Autoimmune Diseases Compared Renal Transplant Recipients J. Am. Soc. Nephrol., March 1, 2003; 14(3): 721 - 727. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
) Episode of an acute rejection. The respective acute rejection episodes are graphically assigned to the nearest PK sampling period preceding the clinical event. The median time period between an acute rejection episode and the preceding PK sampling period was 10 d (range, 0 to 90 d).
) and predose (C12) concentration (
) to discriminate between patients with (n = 12) and without (n = 42) an acute rejection episode during the first 70 d posttransplant. The broken line represents the theoretical ROC curve for no discrimination between the two groups.
) adverse events of leukopenia and/or infection are indicated accordingly. The respective adverse events are graphically assigned to the nearest PK sampling period preceding the event. The median time period between an adverse event and the preceding PK sampling period was 14 d (range, 0 to 38 d).
) to discriminate between patients with (n = 13) and without (n = 41) adverse events of leukopenia (n = 3) and/or infection (n = 10) during the first 70 d posttransplant. The broken line represents the theoretical ROC curve for no discrimination between the two groups.