UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Pharmacokinetics and limited sampling strategies of mycophenolic acid in islet transplant recipients Al-Khatib, Mai 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2009_fall_alkhatib_mai.pdf [ 1.18MB ]
Metadata
JSON: 24-1.0067293.json
JSON-LD: 24-1.0067293-ld.json
RDF/XML (Pretty): 24-1.0067293-rdf.xml
RDF/JSON: 24-1.0067293-rdf.json
Turtle: 24-1.0067293-turtle.txt
N-Triples: 24-1.0067293-rdf-ntriples.txt
Original Record: 24-1.0067293-source.json
Full Text
24-1.0067293-fulltext.txt
Citation
24-1.0067293.ris

Full Text

PHARMACOKINETICS AND LIMITED SAMPLING STRATEGIES OF MYCOPHENOLIC ACID IN ISLET TRANSPLANT RECIPIENTS  by Mai Al-Khatib  B.Sc. Pharmacy, University of Jordan, 2001  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Pharmaceutical Sciences)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) June 2009 © Mai Al-Khatib, 2009 i  Abstract Mycophenolic acid (MPA), the active metabolite of mycophenolate mofetil (MMF) is an immunosuppressant that is used in organ transplantation and  exhibits  wide  inter-patient  variability  in  its  pharmacokinetic  properties in various transplant populations. However, only one study has addressed the pharmacokinetics of MPA in the islet transplant population. The objectives of our study were to characterize the pharmacokinetics of MPA and its two glucuronidated metabolites as well as develop limited sampling strategies (LSSs) for the estimation of MPA area-under- the curve (AUC) in the islet transplant population. Sixteen stable islet transplant recipients on steady-state MPA therapy were recruited. The patients were also on tacrolimus-based, steroid-free immunosuppressant regimens. Blood samples were collected at 0, 0.33, 0.66, 1, 1.5, 2, 3, 4, 6, 8, 10 and 12 hours post-dose. Concentrations of MPA, free MPA, 7-O-mycophenolic acid glucuronide (MPAG) and acylmycophenolic acid glucuronide (AcMPAG) in the plasma samples were measured by a high performance liquid chromatography-ultraviolet detection technique. Conventional pharmacokinetic parameters were determined by non-compartmental analysis. Multiple regression analysis was used to develop the LSSs, using all 16 patients’ profiles. The resulting equations were validated for their bias and precision using the jackknife method. There  was  large  inter-patient  variability  in  all  pharmacokinetic  parameters of MPA, MPAG and AcMPAG. Reasons for this variability are multifactorial and should be the focus of future multicenter studies. Four 3-concentration and one 2-concentration LSS met predetermined criteria and had conventional sampling times. The LSS that we recommend is the one utilizing two concentrations: ii  AUC=1.547+1.417C1+9.448C4. This equation is convenient and can be useful for clinicians in optimizing patient care.  iii  Table of Contents Abstract ...................................................................................... ii Table of Contents ........................................................................ iv List of Tables ............................................................................. viii List of Figures ............................................................................. ix List of Abbreviations ..................................................................... x Acknowledgements ..................................................................... xi Dedication.................................................................................. xii Chapter 1: Introduction 1.1 Islet transplantation ............................................................... 1 1.1.1 Islets of Langerhans ............................................................. 1 1.1.2 History ................................................................................ 1 1.1.3 Indications for islet transplantation ......................................... 2 1.1.4 Success measures ................................................................ 2 1.1.4.1 Procedure success .......................................................... 2 1.1.4.2 Transplantation goals ...................................................... 3 1.1.5 Islet versus pancreatic transplant ........................................... 3 1.1.6 British Columbia program ...................................................... 4 1.1.6.1 Overview ....................................................................... 4 1.1.6.2 Inclusion/exclusion criteria .............................................. 4 1.1.6.3 Islet isolation and infusion process ................................... 5 1.1.6.4 Success rates ................................................................. 6 1.2 Mycophenolate mofetil ............................................................ 9 1.2.1 Overview ............................................................................. 9 1.2.2 History ................................................................................ 9 1.2.3 Physical-chemical properties .................................................. 9 1.2.4 Pharmacology and mechanism of action .................................10 1.2.5 Pharmacokinetics ................................................................10 1.2.5.1 Absorption ....................................................................10 iv  1.2.5.2 Distribution/protein binding ............................................11 1.2.5.3 Metabolism and excretion ...............................................11 1.2.5.4 Enterohepatic recirculation .............................................13 1.2.6 Toxicity ..............................................................................14 1.2.7 Side effects ........................................................................14 1.2.8 Dosing ...............................................................................14 1.2.9 Therapeutic drug monitoring (TDM) .......................................14 1.3 Immunosuppressive regimens in islet transplantation ............... 16 1.3.1 ATG (Atgam®)....................................................................16 1.3.2 Daclizumab (Zenapax®) ......................................................17 1.3.3 Tacrolimus (Prograf®) ......................................................17 1.3.4 Sirolimus (Rapamune®)....................................................18 1.4 UDP-Glucuronosyltransferases (UGTs)..................................... 19 1.4.1 Overview ............................................................................19 1.4.2 Classification ......................................................................20 1.4.3 Mechanism of action ............................................................20 1.4.4 UGTs involved in the metabolism of MPA ................................21 1.4.5 Polymorphism .....................................................................21 1.5 Limited sampling strategies ................................................... 22 1.5.1 Overview ............................................................................22 1.5.2 Methods in establishing limited sampling strategies .................23 1.5.2.1 Bayesian analysis .........................................................23 1.5.2.2 Multiple regression analysis ............................................26 1.5.3 Validation of predictive performance of LSSs ...........................27 1.5.3.1 Overview ......................................................................27 1.5.3.2 Validation approaches ....................................................28 1.6 Significance of research ........................................................ 30 1.7 Objective and specific aims .................................................... 32  v  Chapter 2: Pharmacokinetics of Mycophenolic Acid and its Glucuronidated Metabolites in Islet Transplant Recipients 2.1 Specific aim #1 .................................................................... 33 2.2 Methods .............................................................................. 33 2.2.1 Patient population ...............................................................33 2.2.2 Plasma concentrations of MPA, MPAG , AcMPAG and fMPA ........34 2.2.3 Assessment of pharmacokinetic parameters............................35 2.2.4 Statistical analysis ...............................................................35 2.3 Results ................................................................................ 35 2.3.1 Patient characteristics ..........................................................35 2.3.2 MPA Pharmacokinetics .........................................................37 2.3.3 Free MPA ............................................................................45 2.4 Discussion ........................................................................... 45 2.4.1 Inter-subject variability in pharmacokinetic parameters ...........45 2.4.2 Comparison with the Jacobson study .....................................52 2.4.3 Intra-subject variability ........................................................55 2.4.4 Concomitant immunosuppressants ........................................56 2.4.5 Disease states.....................................................................58 2.5 Summary ..............................................................................59 Chapter 3: Limited Sampling Strategies of Mycophenolic Acid for Estimation of Area under the Concentration-Time Curve in Islet Transplant Recipients 3.1 Specific Aim #2 .................................................................... 60 3.2 Methods .............................................................................. 60 3.2.1 Patient population and MPA concentrations .............................60 3.2.2 Pharmacokinetic parameters assessment ...............................60 3.2.3 Limited sampling strategy determination and validation ...........60 3.2.3.1 The 2-group approach- untransformed data ......................61 3.2.3.2 The 2-group approach, log-transformed data ....................62 3.2.3.3 Jackknife approach ........................................................62 vi  3.2.3.4 Validation of LSS ...........................................................62 3.2.3.5 Comparison with other LSSs ...........................................63 3.3 Results ................................................................................ 63 3.3.1 The 2-group approach..........................................................63 3.3.1.1 Study subjects characteristics .........................................63 3.3.1.2 Untransformed data .......................................................65 3.3.1.3 Log-transformed data ....................................................68 3.3.2 Jackknife approach ..............................................................70 3.3.2.1 Study subjects characteristics .........................................70 3.3.2.3 LSSs using relaxed sampling time criteria (up to 4 hours post-dose) ...............................................................................73 3.3.3 Recommended LSSs ............................................................76 3.3.4 Comparison with other studies ..............................................76 3.4 Discussion ........................................................................... 84 3.4.1 Current status of MPA LSSs ..................................................84 3.4.2 Approach used ....................................................................85 3.5 Summary ............................................................................ 86 Chapter 4: Overall Summary and Conclusion 4.1 Overall discussion and conclusion ........................................... 88 4.2 Strengths and limitations ...................................................... 90 4.3 Current knowledge ............................................................... 91 4.4 Future research .................................................................... 93 5. References 6. Appendices Appendix A: UBC Research Ethics Board's Certificate of Approval ...107 Appendix B: Consent form ........................................................109 Appendix C : HPLC methodology and validation ...........................113 C.1 Plasma concentrations of MPA, MPAG , AcMPAG ................... 113 C.2 Stability of AcMPAG .......................................................... 122 C.3 Free MPA extraction ......................................................... 122 vii  List of Tables Table 2-1: Characteristics of islet transplant patients who participated in this study.....................................................................................36 Table 2-2: PK parameters and metabolic ratios of MPA in islet transplant recipients.....................................................................................37 Table 2-3: Selected MPA pharmacokinetic studies from other research groups.........................................................................................47 Table 3-1: Characteristics of islet transplant recipients in the index and validation groups in randomization A................................................64 Table 3-2: Characteristics of islet transplant recipients in the index and validation groups in randomization B................................................65 Table 3-3: Selected examples of LSSs developed using the 2-group approach with the untransformed data.............................................67 Table 3-4: Selected examples of LSSs developed using the 2-group approach with log- transformed data...............................................69 Table 3-5: Predictive performance of 2- and 3- concentration LSSs (within the 1st 2 hours post-dose) developed using the jackknife method........................................................................................71 Table 3-6: Predictive performance of 2- and 3- concentration LSSs (within the 1st 4 hours post-dose) developed using the jackknife method........................................................................................74 Table 3-7: Selected MPA LSSs established in kidney transplant population from other research groups.............................................77 Table 3-8: Predictive performance of previously published LSSs and results of comparisons in predictive performance between the recommended LSS for islet transplant patients and LSS derived from renal transplant populations............................................................82 Table C-1: Intra-day and inter-day coefficient of variation (CV) of MPA, AcMPAG and MPAG at four concentrations………………………...................115  viii  List of Figures Figure 1-1: The process of islet transplantation................................6 Figure 1-2: Median quarterly HbA1c values for medical vs. islet cell transplant patients........................................................................8 Figure 1-3: Chemical structures of MMF, MPA, MPAG and AcMPAG....12 Figure 1-4: Metabolic pathway for MMF and MPA............................13 Figure 2-1: MPA pharmacokinetic profile (mean ± SD)....................39 Figure 2-2: MPAG pharmacokinetic profile (mean ± SD)..................40 Figure 2-3: AcMPAG pharmacokinetic profile (mean ± SD)...............41 Figure 2-4: MPA pharmacokinetic profile, all 16 patients..................42 Figure 2-5: MPAG pharmacokinetic profile, all 16 patients...............43 Figure 2-6: AcMPAG pharmacokinetic profile, all 16 patients.............44 Figure C-1: Calibration curve of MPA............................................117 Figure C-2: Calibration curve of MPAG..........................................118 Figure C-3: Calibration curve of AcMPAG.......................................119 Figure C-4: HPLC chromatogram of MPA and AcMPAG.....................120 Figure C-5: HPLC chromatogram of MPAG.....................................121  ix  List of Abbreviations AcMPAG:  Acyl mycophenolic acid glucuronide  AUC0-12:  Area-under–the-curve from 0-12 hours post –dose  Cx:  Concentration at time x  DNA:  Deoxyribonucleic acid  fMPA:  Free (concentration of) mycophenolic acid  GI:  Gastrointestinal  HPLC:  High performance liquid chromatography  IMPDH:  Inosine monophosphate dehydrogenase  IS:  Internal standard  LLOQ:  Lowest limit of quantitation  LOD:  Limit of detection  LSS:  Limited sampling strategy  MMF:  Mycophenolate mofetil  MPA:  Mycophenolic acid  MPAG:  7-O-Mycophenolic acid glucuronide  MRA:  Multiple regression analysis  PK:  Pharmacokinetic  RCT:  Randomized controlled trial  TDM:  Therapeutic drug monitoring  x  Acknowledgements I offer my enduring gratitude to my supervisor, Dr. Mary Ensom who has inspired me to pursue my work in this field and supported me each step of the way. She has enlarged my vision of science and always provided rational answers to my endless questions. My experience would have never been as rich without her. Special thanks to Ms. Diane Decarie, Ms. Naomi Johnson, Mr. Wayne Morrissey and Mr. Dwayne Collins for their technical and clinical support. I would like to acknowledge Drs. Nilufar Partovi, Jean Shapiro and Lillian Ting for their support, ideas and advice. Thanks to all my committee members: Drs. Marc Levine, Wayne Riggs, Kishor Wasan and Judy Wong. Your penetrating questions taught me to question more deeply. My deepest gratitude to my family- my husband: Ahmed Barrieshee, my 2 children: Layan and Mahmoud, and all my family back home in Jordan. Your support, encouragement and continuous prayers helped me tackle all the obstacles that came along the way.  xi  Dedication  Gratefully dedicated to my late mother: Mariam Bustami, may God rest her soul in peace.. She raised me to be who I am and would have been very proud to see me...  xii  Chapter 1: Introduction 1.1 Islet transplantation 1.1.1 Islets of Langerhans The Islets of Langerhans are groups of specialized cells in the pancreas that make and secrete hormones. Named after the German pathologist Paul Langerhans (1847-1888), who discovered them in 1869, these cells sit in groups that Langerhans likened to little islands in the pancreas. There are five types of cells in an islet: alpha cells that secrete glucagon, the hormone that raises the level of glucose in the blood; beta cells that secrete insulin; delta cells that secrete somatostatin, an inhibitory hormone affecting the release of numerous other hormones in the body; and PP cells and D1 cells, about which little is known. Degeneration of the insulin-producing beta cells is the main cause of type I (insulindependent) diabetes mellitus (1). The islets of Langerhans will hereon be referred to as ―islet cells‖ or ―islets‖ for short.  1.1.2 History The concept of transplanting pieces or extracts of pancreas in diabetic patients dates back to the 1890s(2). However, it was not until 1972 that Ballinger and Lacy reported that islet isografts from normal rats could successfully reverse chemically-induced diabetes in rats(3). In 1980, the first successful human islet transplant was reported(4). Over the next 3 decades, research in the islet transplant field has focused on two areas: refining islet isolation techniques in order to achieve adequate initial islet mass as well as developing immunosuppressant regimens with low rates  1  of rejection and excellent graft function yet with minimal adverse effects(3). The Edmonton Trial came as a landmark in islet transplantation in 2000(5). The investigators reported islet transplantation in seven type I diabetic  patients.  The  transplants  resulted  in  successful  insulin  independence demonstrated by normalization of glycosylated hemoglobin (a biomarker of glycemic control) and sustained freedom from the need for  exogeneous  insulin.  Their  protocol  included  glucocorticoid-free  immunosuppression combined with infusion of an adequate islet mass. Currently, pancreatic islet transplantation is a promising new area in the treatment of type I diabetes. In North America, 46 medical institutions in the US & Canada have established (or are in the process of establishing) islet transplant program since 1999. Of these, 31 programs have actually performed islet transplantation for a total of 717 infusions in 378 recipients (1).  1.1.3 Indications for islet transplantation Currently, the most common indication for islet transplant is frequent and severe hypoglycemic events. Other indications include clinical and emotional problems associated with the use of exogenous insulin therapy which are severe enough to be incapacitating, and consistent failure of insulin-based management to prevent acute complications(6).  1.1.4 Success measures 1.1.4.1 Procedure success Success of the actual transplantation procedure is usually assessed indirectly by measuring C-peptide levels. However, there is no strategy as yet to monitor rejection. If rejection occurs, it is usually recognized when it is too late to intervene. Signs of rejection are likely an increase in blood 2  glucose readings. Therefore, islet transplant patients are usually put on a higher level of immunosuppression initially to decrease the risk of rejection(7). 1.1.4.2 Transplantation Goals Currently, the major goal of islet transplantation is achieving insulin independence(8). However, this has proved to be quite challenging as it would require several donor pancreata infused over one or more times. Besides, long-term insulin independence can apparently not be achieved with islet transplantation based on the Edmonton protocol(8). A ―more realistic‖ goal, recently suggested, is to convert diabetes from a brittle state into a more easily manageable disease. That would translate clinically into avoidance of severe hypoglycemia and good metabolic control in order to reduce morbidity and mortality resulting from cardiovascular  complications  and  other  chronic  complications  of  diabetes(8).  1.1.5 Islet versus pancreatic transplant Pancreas transplant was the first approach in the biological substitution of beta-cell function in type I diabetes. The first pancreas transplantation was in 1966. From 1966 to 2004, almost 21,000 transplants were performed worldwide. Pancreas transplant is now a well-established clinical indication for patients with type I diabetes also undergoing renal transplantation, with a 3-year insulin independence rate of 80%(6, 8). Graft survival continues to improve. Pancreatic transplant has similar indications to islet transplant and better graft survival rates. However, it carries the additional risks of surgical complications and the risk of vascular complications in patients already affected by advanced vascular disease. These risks are significant enough that one report suggested that mortality among patients undergoing 3  pancreas transplantation alone can be higher than among patients on the waiting list(9).  1.1.6 British Columbia program 1.1.6.1 Overview Here in British Columbia (BC), the Pancreatic Islet Transplant Program started in March 2003 under the guidance and authority of the British Columbia Transplant Society (BCTS). So far (March, 2009), 76 islet transplants have been performed in 31 patients. The program works in collaboration with the Vancouver General Hospital Best Care Diabetes Program to identify suitable recipients. At this time, only patients with type I diabetes without evidence of significant diabetes-related renal disease are being considered for islet transplantation(7). 1.1.6.2 Inclusion/exclusion criteria Inclusion criteria to be a candidate to receive a transplant are(7):   20-65 years of age    Type I diabetes for more than 5 years    Negative/ negligible C-peptide* (fasting and/or stimulated)    Retinopathy- all stages    Renal status:  Creatinine clearance >70 ml/min/1.73 m2  Documented history of albumin/creatinine ratio greater than 1.8 g/mol in men and 2.5 g/mol in women  Normal serum creatinine  * C-peptide is a by-product of insulin production. The level of C-peptide is an estimate of how much insulin is being produced in the body. Exclusion criteria are(7):   Body mass index (BMI) >27 kg/m2    Previous history of myocardial infarction or angina 4    Abnormal baseline MIBI (methoxyisobutyl isonitrile; a radionuclide imaging/cardiac perfusion scan)    Current or recent smoker    Planned pregnancy    Malignant hypertension causing end stage organ damage    Severe concurrent illness likely to limit life or require extensive systemic treatment    Active infection or evidence of ongoing or recurrent viral disease    Inadequate understanding, compliance, or unwillingness to participate with all clinical requirements of the Islet Transplant Program  1.1.6.3 Islet isolation and infusion process Islets for transplant are obtained from the pancreas of a cadaveric donor, typically someone older than 50 years of age and with a BMI greater than 30. This is different from suitable organs for whole pancreas transplants, which typically come from younger, more fit donors. Thus, there is no competition between whole and islet pancreas transplants for organs(7). Once human pancreata are obtained from cadaver organ donors, they are subjected to islet isolation. Briefly, the isolation process includes intraductal collagenase (liberase) perfusion, continuous digestion, and density gradient purification. Recently, an extra step was added to the protocol. Impure tissue fraction (i.e., usually a useless leftover from the isolation process) is further cultured in vitro and then repurified to retrieve additional islets(10). Purified islets are then implanted by ―percutaneous portal embolization‖: a procedure in which a radiologist uses ultrasound and radiography to guide placement of a catheter through the upper abdomen and into the portal vein of the liver of the recipient. The islets are then infused through the catheter into the liver, providing more than 10,000 islet 5  equivalents (IE) per kilogram of body weight. The patient is usually under local anesthesia; however, general anesthesia can sometimes be used if the patient cannot tolerate local anesthesia. The process is usually achieved by infusions from one to three donors per patient(10). Newly transplanted islets take some time to attach to new blood vessels but usually begin releasing insulin within hours of infusion. The patient’s exogenous insulin needs typically start to decline in the few days to weeks following the infusion and  insulin independence is quickly  achieved(11). Figure 1-1 illustrates the transplantation process. Figure 1-1: The process of islet transplantation (12) (illustration by Giovanni Maki)  1.1.6.4 Success rates In a 2008 multi-year analysis study of the islet transplantation program in BC, the program defined the goal of islet transplantation: to achieve 6  and maintain insulin independence as long as possible. Patients who achieved initial insulin independence for greater than 1 month but later required insulin were treated with supplemental donor islets(11). The report included data up to July 15, 2008 (3-year follow-up). Of 25 patients who have completed their transplant protocol, 16 remained insulin- independent by the time of the report. Patients with partial islet graft function who have resumed insulin were taking 33%-75% of their pre-transplantation dose. The report compared islet transplantation with intensive medical therapy on progression of complications in type I diabetes. The metabolic outcomes  looked  at  were  glycosylated  hemoglobin*  (HbA1c),  nephropathy, retinopathy, and neuropathy. The authors concluded that islet transplant provided improved glycemic control (and lessened the progression of diabetic retinopathy(11). Figure 1-2 illustrates median quarterly HbA1C for medical versus islet transplant subjects during the study follow-up. *Glycosylated hemoglobin (HbA1c) is a form of hemoglobin used primarily to identify the average plasma glucose concentration over prolonged periods of time.  7  Figure 1-2: Median quarterly HbA1c values for medical (MED) vs. islet cell transplant (ICT) patients and corresponding intraquartile range. Warnock et al. (11)  Measurements for subjects were combined for each quarter of follow-up and the median value plotted as shown. At all time periods studied, HbA1C (%) was lower for ICT and pooling all numbers to calculate total glycemic exposure during the study period, HbA1C was 7.4 for medical versus 6.6 for ICT (P<0.01)(11).  8  1.2 Mycophenolate mofetil 1.2.1 Overview Mycophenolate  mofetil  (MMF)  (Cellcept®)  is  a  standard  immunosuppressant used in solid organ transplant to prevent acute rejection. MMF is a prodrug with mycophenolic acid (MPA) being the active metabolite. Its clinical utility as an immunosuppressant was realized in the mid-1990s when it was approved for prophylactic use against acute graft rejection in renal transplant patients. Currently, its use has extended to other transplanted organs like the liver, lung and heart and it is a mainstay of islet transplants as well(13, 14).  1.2.2 History MMF was developed by Nelson, Allison and Eugui of Roche Laboratories (formerly known as Syntex Research) with the aim to provide a new immunosuppressant that has selective and reversible antiproliferative effects(14-16)(17). MPA was first isolated as a fermentation product from several Penicillium species found in corn mold in 1898. The ester derivative of MPA, i.e. MMF, was found to have better bioavailability than MPA due to better solubility properties(18).  1.2.3 Physical-chemical properties Mycophenolate mofetil is a white to off-white crystalline powder. The chemical  name  of  MMF  is  2-morpholinoethyl  (E)-6-(1,3-dihydro-4-  hydroxy-6-methoxy-7-methyl-3-oxo-5-isobenzofuranyl)-4-methyl-4hexenoate. It has a molecular mass of 433.5 and an empirical formula of C23H31NO7. MMF is slightly soluble in water (43 μg/mL at pH 7.4); the solubility increases in acidic medium (4.27 mg/mL at pH 3.6). It is freely soluble in dimethyl sulfoxide, tetrahydrofuran, acetone, acetonitrile, 9  dichloromethane, and ethyl acetate; soluble in methanol and propylene carbonate; sparingly soluble in anhydrous ethanol; slightly soluble in isopropanol and diethyl ether; and very slightly soluble in hexane(19). The chemical structure of MMF is presented in Figure 1-3.  1.2.4 Pharmacology and mechanism of action MPA is a selective, reversible and non-competitive inhibitor of inosine monophosphate dehydrogenase 2 (IMPDH 2). IMPDH 2 is the ratelimiting enzyme in the de novo pathway of purine synthesis. It is abundant in proliferating T and B lymphocytes(17). As a result, DNA synthesis is blocked and proliferation of T and B lymphocytes in response to antigen stimulation is markedly inhibited in the presence of MPA(20). MMF works as a selective immunosuppressant because the T and B lymphocytes are heavily dependent on the de novo pathway for purine synthesis, while other cells can utilize the salvage pathway. The salvage pathway is one in which nucleotides (purine and pyrimidine) are synthesized  from  intermediates  in  the  degradative  pathway  for  nucleotides. In addition, IMPDH 2 is five times more susceptible to inhibition by MPA than IMPDH 1. While IMPDH 1 is found in abundance in the kidney, intestine, spleen and other body organs, high levels of IMPDH 2 are expressed in actively proliferating lymphocytes thus adding to selective inhibitory properties of MPA(14, 16).  1.2.5 Pharmacokinetics 1.2.5.1 Absorption The absorption of MMF from the gastrointestinal tract is > 90%(21). Once in the blood, it undergoes instantaneous and complete hydrolysis by serum carboxylesterases to give MPA. MPA maximum concentration (Cmax) is usually achieved within 1 hour. Fatty food seems to affect the 10  rate but not the extent of absorption. As such, area under the plasma drug concentration-time curve (AUC) is comparable in the fed versus overnight fasting state. However, time to maximal concentration (Tmax) is slightly delayed and maximal concentration (Cmax) is decreased by 25% in fed vs. fasting state(22). 1.2.5.2 Distribution/protein binding In plasma, MPA is highly protein bound. It binds extensively to albumin with a bound fraction of approximately 97-99%. MPA does not bind significantly to α1-acid glycoprotein(23). The main metabolite of MPA, 7O-glucuronide (MPAG), is also highly protein bound, about 82%(19). The apparent volume of distribution of MPA is estimated at 4 L/Kg in healthy volunteers(19). 1.2.5.3 Metabolism and excretion MPA is eliminated mainly by glucuronidation catalyzed by a group of phase II metabolizing enzymes: UDP-glucuronosyltransferases or (UGTs). Approximately 87-94% of the MPA appears in the urine as the inactive 7O-glucuronide MPAG (24). The rest of the conjugation products are accounted for by an acyl glucuronide (AcMPAG), and two glucoside conjugates (mycophenolate 7-O-glucoside: MPAGI and mycophenolate acyl glucoside: AcMPAGI). Only 6% of MPA is excreted in the feces in an unchanged form(24). The enzymes UGT1A9 and 2B7 are believed to be the major  isoforms involved in conjugating MPA. These 2 forms have  high hepatic and renal expression. In vitro experiments show that UGT1A9 is responsible for producing 55%, 75% and 50% of MPAG in the liver, kidney and intestinal mucosa, respectively(25). MPAG is also formed by UGT 1A7, 1A8 and 1A10 (found in the kidney and intestine); UGT 2B7 is the only isoform reported so far to produce AcMPAG.  11  Recently, a minor phase I metabolite has been identified as 6-Odesmethyl MPA (DM-MPA). Cytochrome P450 3A isoforms are believed to be mainly involved in its production(23). Estimates of mean elimination half-life of MPA range from 9-17 hours(26-28). Figure 1-3 depicts the chemical structure of MMF, MPA, MPAG and AcMPAG and Figure 1-4 summarizes the metabolic pathway of MMF and MPA. Figure 1-3 : Chemical structures of MMF, MPA, MPAG and AcMPAG  MMF  MPA  MPAG  AcMPAG  12  Figure 1-4: Metabolic pathway for MMF and MPA Shaw et al. (28) EHC: enterohepatic circulation  1.2.5.4 Enterohepatic recirculation MPA also exhibits prominent features of enterohepatic circulation. MPAG, conjugated in the liver, is excreted through the biliary system via multidrug resistance protein 2 (MRP2; ABCC2) - mediated transport into the intestine, where it undergoes hydrolysis to MPA and subsequent reabsorption. That is reflected by secondary peaks appearing in the plasma MPA concentration – time profiles anywhere from 4 to 12 hours following the morning dose of MMF. The contribution of enterohepatic circulation to the AUC has been estimated at approximately 37%, ranging from 10 to  13  61%(23, 29).  The organic anion transport protein, MRP2, is expressed  mainly in membranous tissues of bile duct, kidney and intestine(24).  1.2.6 Toxicity Contrary to the inactive MPAG metabolite, AcMPAG has been reported to have a pharmacological activity similar to MPA(29). Acyl glucuronides are also known for their toxicity. They are reactive electrophilic metabolites that can bind covalently to proteins, lipids and nucleic acids and result in direct tissue damage. This is postulated to be one mechanism explaining the adverse effects (e.g. gastrointestinal toxicity and leucopenia) seen with mycophenolate mofetil therapy. The MPA glucosides do not exhibit immunosuppressive activity and little is known about their toxicity(21).  1.2.7 Side effects MPA is fairly well tolerated(30). However, common adverse events of MPA are usually gastrointestinal and hematologic. Gastrointestinal effects include gastrointestinal upset (nausea, vomiting, mild diarrhea, ulcer), and hematologic effects include leucopenia, anemia and increased risk of infections. Other reported side effects include: headache, mild weakness dizziness or tremor, insomnia and swelling of the lower legs or feet. There is also an increased risk of lymphoma and other cancers(19).  1.2.8 Dosing MMF is usually prescribed at a fixed dose given twice daily. The total daily dose varies from one transplant population to another but ranges from 1.5 to 3 gm(19).  1.2.9 Therapeutic drug monitoring (TDM) Pharmacokinetic-pharmacodynamic studies have shown a correlation between MPA concentration and pharmacodynamic effects(28, 31). Based on these studies, it is reasonable to use MPA AUC as a surrogate marker 14  for MPA effects. However, whether monitoring concentrations of MPA in plasma of transplant patients will result in a lower incidence of graft rejection or side effects remains unclear as the published evidence so far is still inconclusive(30). Only two published randomized controlled trials (RCTs) have done a direct head-to-head comparison of monitoring vs. no monitoring (32)(33) The first study(32) compared (over a period of 12 months) kidney transplant patients who were randomly assigned to either a fixed dose of MMF that could be clinically adjusted or a ―concentration-controlled‖ dose that was adjusted to achieve an MPA set target level. There were more treatment failures (i.e. acute rejection, loss of graph or death) in the fixed dose group than the concentration-controlled group. The second, more recent RCT (33) compared in a similar manner 2 groups of kidney transplant patients either on a fixed dose MMF or clinically adjusted dose but reported dissimilar results. Over a 12- month follow-up period, these authors observed no statistically significant difference in the incidence of treatment failure (which was a composite of biopsy-proven acute rejection, graft loss, death or MMF discontinuation). However, they also reported that MMF dose adjustments based on target MPA exposure was not successful, partly because physicians seemed non-adherent to required dose increments and rather reluctant to implement substantial dose changes. Two published observational studies (in 3 papers)(34-36) provided conflicting results about the value of monitoring MPA levels. Meiser et al.(34, 35) compared 2 groups of heart transplant recipients. Group one (n=15) was on a fixed MMF dose while group two (n=30) had their MMF dose adjusted according to a target pre-dose plasma concentration. Over the follow-up period, 27 patients of group two remained rejection-free while only 5 group one patients remained rejection-free over the follow15  up period (mean length: 696 days and 436 days for groups one and two, respectively). On the other hand, Flechner et al.(36) conducted a similar study with a bigger sample size (n=160 and 100 for fixed dose and concentration-controlled  dose  groups,  respectively)  and  found  no  differences between the groups in acute rejection (confirmed by biopsy) as well as no difference in occurrence of viral infection. Until more conclusive evidence about the value of routinely measuring MPA levels is established, TDM for MPA should be decided on a case by case basis(30).  1.3 Immunosuppressive regimens in islet transplantation Immunosuppressive regimens in islet transplantation usually consist of induction  followed  by  maintenance  immunosuppression.  Induction  immunosuppressants are usually given shortly before and/or in the immediate  few  days  following  the  transplantation  procedure.  Antithymocyte globulin (ATG) and daclizumab are the most commonly used immunosuppressants(5, 11). On the other hand, maintenance immunosuppressants are started shortly after transplantation and are usually prescribed for life. Regimens are usually center specific and tailored to patient’s tolerance. The 3 most commonly  prescribed  immunosuppressants  are  tacrolimus  and/or  sirolimus and/or mycophenolate(5, 10). Contrary to other transplant populations’ immunosuppressive regimens, islet transplant regimens are glucocorticoid-free(5, 10).  1.3.1 ATG (Atgam®) ATG is a polyclonal antibody that selectively destroys T lymphocytes. It is the gamma globulin fraction of antiserum from horses that have been 16  immunized against human thymocytes. Its mechanism of action is not clearly understood; however, it may involve elimination of antigenreactive T lymphocytes in peripheral blood or alteration of T-cell function(37).  1.3.2 Daclizumab (Zenapax®) Daclizumab is an IgG1 humanized monoclonal antibody produced by recombinant DNA technology that specifically binds to the alpha subunit of the IL-2 receptor expressed on the surface of activated T cells. When administered,  daclizumab  inhibits  IL-2  mediated  activation  of  lymphocytes, which is a critical pathway in cellular immune response involved in the rejection process(37).  1.3.3 Tacrolimus (Prograf®) Tacrolimus is a macrolide immunosuppressant produced by the fungus Streptomyces tsukubaensis. Tacrolimus inhibits T-lymphocyte activation. Although the exact mechanism of action is not known, experimental evidence suggests that tacrolimus binds to an intracellular protein, FKBP12.  A  complex  of  tacrolimus-FKBP-12,  calcium,  calmodulin,  and  calcineurin is then formed and results in inhibition of calcineurin phosphatase  activity.  This  effect  may  be  responsible  for  the  dephosphorylation and translocation of nuclear factor of activated T-cells (NF-AT), a nuclear component thought to initiate gene transcription for the formation of lymphokines (such as interleukin-2, gamma interferon). The net result is the inhibition of T-lymphocyte activation (38). Tacrolimus is available in intravenous as well as oral (0.5, 1, 5 mg) formulations. Oral absorption of tacrolimus is incomplete and variable. Oral bioavailability is around 18% in healthy volunteers and ranges from 17- 23% in various transplant populations. Food also impairs the absorption of tacrolimus (38). 17  The plasma protein binding of tacrolimus is approximately 99%. Tacrolimus is bound mainly to albumin and alpha-1-acid glycoprotein, and is also highly associated with erythrocytes. The mean ratio of whole blood concentration to plasma concentration of tacrolimus is 35 (range 12 to 67). Due to intersubject variability in tacrolimus pharmacokinetics, individualization of dosing regimen is necessary for optimal therapy(38). Tacrolimus is extensively metabolized by the cytochrome P-450 system (CYP3A). It is therefore susceptible to pharmacokinetic drug-drug interactions by CYP3A inducers and inhibitors. One in vitro study(39) showed that tacrolimus is an effective inhibitor of UGT enzymes; thus, it was suggested that tacrolimus co-administration would result in reduced formation of MPAG and lead to higher MPA concentrations (27, 40). Further discussion of the MPA-tacrolimus interaction and whether this can potentially be clinically significant is included in section 2.4.4, Chapter 2. Two potentially significant adverse events of tacrolimus in the islet transplant population are insulin-dependent post-transplant diabetes mellitus  (PTDM)  and  dyslipidemias:  hypertriglyceridemia  and  hypercholesterolemia(38). As many islet transplant patients already have or are at risk of developing dyslipidemia (because of their current or previous diabetic condition) this adverse event may potentially be clinically significant. However, when compared to cyclosporine, which also works as a calcineurin inhibitor, tacrolimus has a better profile on lipid metabolism(41) as well as superiority in improving graft survival and preventing acute rejection(42).  1.3.4 Sirolimus (Rapamune®) Sirolimus is a macrocyclic lactone immunosuppressant produced by the fungus Streptomyces hygroscopicus. Sirolimus inhibits T-lymphocyte activation and proliferation by a mechanism that is distinct from that of other immunosuppressants. Sirolimus also inhibits antibody production. 18  In cells, sirolimus binds to the immunophilin, FK Binding Protein-12 (FKBP-12), to generate an immunosuppressive complex. However, unlike tacrolimus, the sirolimus: FKBP-12 complex has no effect on calcineurin activity. Rather, this complex binds to and inhibits the activation of a specific cell cycle regulatory protein called the mammalian target of rapamycin (mTOR). The mTOR is a key regulatory kinase and its inhibition suppresses cytokine-driven T-cell proliferation(43). Sirolimus is only available orally (oral solution: 1mg/mL; tablets: 1 , 2 and 5 mg). Systemic availability of sirolimus is approximately 14% when administered as an oral solution. When administered as tablets, the mean bioavailability is about 22% higher (relative to the oral solution). Concomitant food intake may alter the bioavailability of sirolimus (oral solution or tablet). Thus, sirolimus should be taken consistently, either with or without food in order to minimize blood level variability. Sirolimus is extensively metabolized by the CYP3A4 isozyme in the gut wall and liver and undergoes counter-transport from enterocytes of the small intestine  by  the  P-glycoprotein  drug  efflux  pump.  Consequently,  absorption and elimination of systemically absorbed sirolimus may be influenced by drugs that affect these proteins(43).  1.4 UDP-Glucuronosyltransferases (UGTs) 1.4.1 Overview UDP-Glucuronosyltransferases, known as UGTs, are a class of phase II metabolizing  enzymes.  Phase  II  metabolism  generally  involves  conjugation of lipophilic moieties to hydrophilic compounds so that the resulting metabolites have increased water solubility and thus are easier to eliminate in the bile or urine.  19  Phase II conjugation reactions include sulfation (conjugation of a sulfate), acetylation (conjugation of an acetyl group), methylation (conjugation of a methyl group), glucuronidation (conjugation of glucuronic acid) and conjugation with amino acids. UGT substrates are comprised of a wide range of endogenous as well as exogenous compounds including androgens, estrogens, bilirubin, morphine, acetaminophen, salicylates and mycophenolate(44). In fact, approximately one-third of drugs undergoing phase II metabolism are conjugated by UGTs(45). UGTs are mainly expressed in the liver, kidney and gastrointestinal tract. However, they are also found in other body organs like the brain, uterus and breast(45).  1.4.2 Classification To date, as many as 19 forms of UGT have been identified in humans. Based on their sequence analysis, they can be divided into two families: UGT1 and UGT2. The UGT1 family is less than 50% similar in sequence to the UGT2 family. All the UGT1 forms have identical carboxy-terminal domains and are encoded by a single long gene locus extending more than 100 kilobases on chromosome 2. On the other hand, UGT2 forms are encoded by separate genes on chromosome 4 and are divided into 2 subfamilies; 2A and 2B. Of the 19 UGT sequences identified, three appear to encode nonfunctional proteins (UGT1A2 UGT1A11, UGT1A12). Substrate specificities for most of the UGT forms have been characterized in cDNA expression systems(46).  1.4.3 Mechanism of action It is believed that UGT enzymes function as dimers transferring glucuronic acid from UDP-glucuronic acid to hydroxyl, carboxyl, amino, 20  thiol and carbonyl groups on a vast array of lipophilic chemicals rendering them more water soluble and thus more readily excreted in the urine or bile.  Glucuronidation  can  also  occur  at  acetyl  groups  (e.g.  in  acetylsalicylic acid, mycophenolic acid) to form an acyl glucuronide. Acetyl metabolites can be reactive and potentially toxic(29).  1.4.4 UGTs involved in the metabolism of MPA MPA glucuronidation has been shown to occur in the liver and the kidney as well as in cell lines derived from the gastrointestinal tract. The liver is considered the major organ of MPA glucuronidation. However, the kidney has also been reported to have as high as 80-fold glucuronidation activity compared to the liver. The intestine, on the other hand, has considerably less glucuronidation activity compared to both other organs(46). The specific enzymes involved in the glucuronidation of MPA have been discussed in section 1.2.5.3 of this Chapter.  1.4.5 Polymorphism Genetic polymorphisms have been identified for the following UGT enzymes: UGT1A1, UGT1A3, UGT1A4, UGT1A6, UGT1A7, UGT1A8, UGT1A9, UGT2A1, UGT2B4, UGT2B7, UGT2B15, and UGT2B28(47). Disequilibrium between the allelic variants of UGT1A1, UGT1A6, and UGT1A7 and the co-occurrence of the UGT1A1*28, UGT1A6*2, and UGT1A7*3 alleles have also been reported(47). Some of these genetic polymorphisms have led to clinically significant alterations in enzyme activity. One example is UGT1A1, which is a major conjugating enzyme responsible for the homeostasis of bilirubin and the glucuronidation of a wide selection of xenobiotics. A common genetic polymorphism in the promoter region of the UGT1A1 gene, denoted as UGT1A1*28, leads to reduced enzyme expression and is associated with 21  Gilbert's syndrome(48, 49). The UGT1A1*28 genotype is also associated with reduced elimination of SN-38, the active metabolite of the chemotherapeutic  agent  irinotecan.  It  has  been  suggested  that  prospective screening of this polymorphic allele may decrease the incidence  of  irinotecan-associated  toxicity  through  permitting  an  alternative therapy to be initiated(50). One other interesting UGT enzyme is UGT2B7. It is the major enzyme responsible for the glucuronidation of morphine to produce morphine-3O-glucuronide (M-3-G) and morphine-6-O-glucuronide (M-6-G)(51). The A to T transversion at nucleotide 802 leads to a change in amino acid sequence, H268Y. This allele is designated as UGT2B7*2 and one-third of the Caucasian population expresses the homozygous genotype(45, 52, 53). However, it is not clear yet if expressing this genotype actually results in clinically significant differences in the metabolic ratios for morphine(47).  1.5 Limited sampling strategies 1.5.1 Overview The AUC is the most commonly used parameter to characterize total body exposure to a medication. However, in order to calculate the AUC, often more than 10 samples are required over a dosing interval. This is impractical for clinical purposes as it is costly, time consuming and uncomfortable for patients. Measurement of a single trough concentration (C0) may reflect total drug exposure when the drug has consistent bioavailability and elimination properties. Unfortunately, this is unsuitable for drugs that have variable and unpredictable characteristics as is the case with MPA. One possible approach commonly used is to predict the AUC from a limited blood sampling schedule. This approach is commonly 22  known as a limited sampling strategy (LSS) and is usually achieved with 3 or fewer plasma drug concentrations(54). In addition to TDM, which aims at individualizing drug therapy, limited sampling strategies have other applications in other clinical and research settings. One such example is bioequivalence studies. These studies involve comparison of two drug formulations to assess for differences in their pharmacokinetic parameters, usually AUC, Cmax and Tmax. LSSs can be helpful in limiting the number of blood draws per subject when estimating these parameters. Another example is to estimate the AUC of drugs used as in vivo probe substrates for the evaluation of enzyme activities. The AUC of the probe substrate usually reflects the activity of the specific enzyme involved in its metabolism. Applying an LSS can simplify the pharmacokinetic profiling and AUC calculation(55). LSSs for MPA have been suggested by numerous groups(56-67). These studies developed the LSSs largely in the kidney transplant population. LSSs have also been developed for other transplant populations like the heart, lung and liver(54, 68, 69). However, to date, there is no LSS established for the islet transplant population. An LSS established in a specific population is not always suitable for other populations especially when the transplanted organ (e.g. the kidney) is involved in the disposition of the drug in question.  1.5.2 Methods in establishing limited sampling strategies There are 2 main approaches to developing limited sampling models, Bayesian analysis and multiple regression analysis: 1.5.2.1 Bayesian analysis The Bayesian approach applies Bayes theory to predict individual AUC values. It requires both population data as well as individual data(55).  23  In brief, plasma concentration profiles as well as demographic data are obtained from patients and then incorporated into a Bayesian software program  (e.g.  NONMEM,  Adapt  II)  that  has  typical  population  pharmacokinetic (PK) parameters for commonly used drugs already stored in it. The Bayesian approach utilizes both types of data (i.e. population and individual) to predict individual PK parameters. As more individual data are gathered (e.g. multiple profiles, demographic data), stored population data contribute less to the overall prediction and the parameter in question becomes more individualized. To develop an LSS using the Bayesian approach, individual as well as population data are required. If population information is unavailable, the software can utilize a part of its dataset (index set) to determine the a priori parameters. The prediction of parameters is achieved by minimizing the Bayesian function:  Bayesian Function = ∑  (Ppop-P̂ )2 + ∑ var(P)  (Cobs- Ĉ )2 var(C)  where Ppop is the population average of parameter P; P̂ is the individual expected average of parameter P; var(P) is the variance of the estimated parameter P; Cobs is the observed concentration value, Ĉ is the predicted concentration value, and var(C) is the variance of the predicted concentration(55). Validation of the Bayesian model is essential and is discussed in further detail in section 1.4.3 of this Chapter. LSSs  developed  with  the  Bayesian  advantages:  24  approach  have  the  following    Sampling time can be flexible. This is helpful in clinical settings where specific time samples cannot be guaranteed.    The system is dynamic: new data can be incorporated all the time to refine prediction of PK parameters. Newly-acquired population data as well as individual data (e.g. demographic data: lifestyle, age, comedications) can be included to provide a better estimate of PK parameters.    More than one PK parameter can be calculated simultaneously.    When incorporated within PK software programs, it can be useful in providing visual comparisons as well as simulations for predicted values and plasma concentration profiles.  However, the Bayesian approach has the following limitations:   It runs complicated calculations and algorithms. These usually require specialized software programs that have a high cost of setting, running and training health care personnel to use properly.    A pharmacokinetic model needs to be specified for the drug (e.g. one compartment,  2  compartment,  etc).  Parameter  prediction  is  dependent on how well the model fits the drug. This can be problematic if no model has yet been assigned for the drug or if more than one model has been specified in the literature.   Finally, the Bayesian approach relies on population parameters for precise estimates as mentioned above. If population estimates are not available, the index set of profiles can be used in their place, but the amount of uncertainty in prediction can be large especially if the sample size of the validation group is small.  25  1.5.2.2 Multiple regression analysis Multiple regression analysis (MRA) correlates a dependent variable (e.g. AUC) to independent variables (e.g. the plasma concentrations at different time points) via stepwise regression analysis. The resulting relationship is expressed as a function in the following form: AUC= b+ M1Ct1+ M2Ct2+…MiCti, where b is the y-intercept constant, Ct1,Ct2…Cti are the concentrations obtained at times t1, t2…ti etc and M1, M2 ..Mi are fitted constants at each time point. MRA has a number of advantages that makes it a popular method for developing LSSs(70):   MRA is simple to use. Most statistical software programs can perform MRA and the resulting LSS equations can be easily applied with little pharmacokinetic background. This makes it readily applicable in the clinical setting.    It is model independent. No pharmacokinetic information about the drug is required in order to develop the LSSs.  On the other hand, MRA comes with some limitations including(55):   MRA depends on timed concentrations. If deviations occur in the sampling time, they can greatly affect the accuracy of the predicted AUC especially if the LSS equation utilizes early post-dose sampling times when drug concentration tends to change rapidly.    The developed LSSs are restricted to the dosing regimen as well as the population in which it was developed. Its accuracy cannot be guaranteed if used for other dosing regimens or in a different subpopulation.  26  1.5.3 Validation of predictive performance of LSSs 1.5.3.1 Overview Once developed, LSSs (derived by either way) need to be validated for their predictive performance. This is usually achieved by testing them on another set of patients’ concentration-time profiles. The predicted parameter value (i.e. the AUC) is then compared to the observed value. No set rules regarding the evaluation of the prediction are available. There are, however, widely accepted guidelines suggested by Sheiner and Beal for testing predictive performance(71). The 2 main criteria for assessment of the LSSs’ predictive performance are bias (systemic error) and precision (random/sampling error). Absolute bias can be measured by the mean prediction error (ME), while absolute precision is measured by the root mean squared prediction error (RMSE) or, alternatively, the mean absolute error (MAE): Equation (1):  ME = 1/N Σ(Pei)  Equation (2):  RMSE =√1/N Σ(Pei)2  Equation (3):  MAE = 1/N Σ|Pei|  Where Pe = prediction error= predicted value – actual value N = the number of data points Relative bias and precision can be easily calculated by converting absolute bias and precision into percentages(55). When MRA is used to develop the LSS, the correlation coefficient of each developed equation (r) or the coefficient of determination (r2) show how well the predicted values correlate to the observed value (by convention, a value of ≥0.7 is considered acceptable in MPA LSSs)(55). However, this is merely a demonstration of association that provides no information as 27  to the bias or precision of the prediction. To illustrate, the predicted values may consistently be two times higher than the observed values but still have excellent correlation. Therefore, studies that develop LSSs depending only on r or r2 with no further validation should be interpreted with caution (55). 1.5.3.2 Validation approaches As mentioned above, in order for the predictive performance to be reliable, the developed LSSs should be tested on a different set of patients than those used to develop it. This can be achieved by more than one approach depending on the data available. 1.5.3.2.a 2-group approach The most widely used approach is to (randomly) split the available plasma profiles into 2 groups: the index set and the validation set. The index set data are used to develop the limited sampling model, whereas the validation set is used to test and assess the predictive performance of the developed LSS model. Although easy and straightforward to apply, this approach uses only part (usually around half) of the profiles to generate each model. While this may not be a limitation if the number of plasma profiles available is relatively large and variance is small, an approach that makes use of all available profiles to generate the model may be better suited when the number of available profiles is limited and wide variability is observed. 1.5.3.2.b Jackknife approach The jackknife method is a resampling scheme in which the LSS equation is generated N number of times, where N is the number of patients’ plasma profiles available. Each time, an LSS is derived from N-1 patients and used to predict the dependent variable (i.e. AUC) of the Nth patient.  28  Thus, a slightly different model is used to predict the AUC of each patient and then to test the predictive performance(65, 72). This technique has considerable theoretical advantages when it is used to analyze  small  expensive-to-collect  data  sets  where  distributional  assumptions are usually unclear and further data collection may be difficult to achieve(65). However, it is less widely used as it tends to be more tedious because each model has to be generated and validated N number of times compared to only once in the 2-group approach. 1.5.3.2.c Bootstrapping approach The bootstrap method generally uses an algorithm to estimate confidence intervals for pharmacokinetic parameters. It involves repeated random sampling of subjects in the database, with replacement of the original dataset by another dataset of the same size with different combination of subjects. As the number of bootstrap approaches infinity, the sample standard deviations for the parameters approaches the ―true‖ (but unknown) standard deviations(68). The name may come from the phrase ―pull up by your own bootstraps‖ which means ‖rely on your own resources‖(73). When in the context of validating LSS, the bootstrapping approach can be achieved briefly as follows(74):   A large number of samples (usually a few hundred) are drawn with replacement from the original set of profiles. For each of these samples, a linear regression model is calculated.    Final model coefficients are determined as the mean of each corresponding coefficients derived from all samples    The accuracy and precision for all models are then derived as the difference/absolute difference between predicted and measured AUC, respectively. 29  The bootstrap resampling method is a robust and well-characterized method that makes efficient use of available data(74). However, its use is limited by being time-consuming and computer-intensive.  1.6 Significance of research Dosing information for MMF in islet transplant recipients is almost exclusively extrapolated from the renal transplant literature. Patients are currently discharged on a 1gm twice a day dose (i.e. the current recommended dosage for the kidney transplant population)(7). However, when recruited for study participation, they were usually at a lower dose, related to adjustment as necessitated by side effects (commonly GI intolerance or lower leucocyte count). Islet transplant immunosuppressive regimens differ from those for other transplant types in that they are steroid-free. Corticosteroids are shown to induce UGT enzymes, the main metabolic pathway of MPA(27). Direct studies on pharmacokinetic interactions between corticosteroids and MPA are limited as well as data regarding the impact of corticosteroids on those specific UGTs involved in MPA conjugation. However, there is evidence that steroids interfere with MPA bioavailability in solid organ transplant recipients and changes in MPA concentrations may be clinically relevant(75). Islet transplant recipients are also unique in their disease state. That is, diabetes is known to influence both the pharmacokinetics and dynamics of drugs. Diabetes-related delayed gastric emptying may affect drug absorption and bioavailability. Diabetes could also alter the metabolism of drugs by affecting the amount or activity of metabolizing enzymes(21). One prospective study compared the 12-hour trough levels of MPA between diabetic and non-diabetic kidney transplant recipients and found 30  a statistically significant difference between the 2 groups (non-diabetic patients had almost double the mean trough level of diabetics)(76). Another case-controlled study compared the concentration – time profiles for MPA, free MPA concentration, MPAG and AcMPAG in diabetic vs. nondiabetic kidney recipients. No significant difference was observed in the characteristics of MPA or fMPAconcentrations of fMPA or the metabolites between the 2 groups(20). However, there was an initial slower absorption rate. Despite that, only one pharmacokinetic study(77) on MPA has been published in the islet transplant subpopulation. The previous study(77) provided useful pharmacokinetic information for total and free MPA as well as MPAG over a one-year follow-up period. The authors reported a large inter- as well as intra- patient variability of total MPA pharmacokinetic measures for Cmax, trough concentrations and AUC over the 12h dosing interval (AUC0-12). However, this study was limited by its small sample size (n=8) and specific population: all 8 patients were females who had received only one transplant and achieved insulin independence. The study was also limited by lack of information on the disposition of AcMPAG, which unlike MPAG, has pharmacologic and proinflammatory activity(24). In contrast, our current islet recipient population (in British Columbia) includes both males and females. Patients have usually received between 1- 4 transplants and have achieved either total or partial insulin independence. Unlike the previous study(77), we measured AcMPAG concentrations (in addition to total and free MPA as well as MPAG). Also, to our knowledge, no published study has developed or validated LSSs of MPA in the islet transplant population. The ramifications of underdosing MMF include recurrent/recalcitrant acute rejection, a major risk factor for the development of chronic rejection, which virtually always portends impaired quality of life and graft survival rates.  On  the  other  side  are  serious 31  adverse  effects  including  gastrointestinal toxicity, bone marrow suppression, and severe infections. Thus, the pharmacokinetics information generated from our study is expected to be helpful in effective management of islet transplant patients. Study results will be incorporated immediately into patient care to optimize dosing for mycophenolate in this important but little-studied patient subpopulation.  1.7 Objective and specific aims The objective of this study was to address the following two specific aims: 1. To characterize the pharmacokinetic properties and metabolic ratios of MPA in stable islet transplant recipients. 2. To establish clinically convenient limited sampling strategies for the prediction of MPA AUC in the islet transplant population.  32  Chapter 2: Pharmacokinetics of Mycophenolic Acid and its Glucuronidated Metabolites in Islet Transplant Recipients 2.1 Specific aim #1 The objective of this part of the study was to characterize the pharmacokinetic properties and metabolic ratios of MPA in stable islet transplant recipients.  2.2 Methods 2.2.1 Patient population This was a single prospective (2-year), open-label clinical study involving the same subject population. Patients were recruited from the islet transplant program at Vancouver General Hospital and Health Centre. Patient inclusion criteria were islet transplant patients who are: on a ―steady-state‖ twice-daily dosage of MMF (attainment of steady state assumed when patients have taken MMF for at least 5 days without a dosage change); at least 16 years of age; and able to provide informed consent. Exclusion criteria were subjects who are: refusing or unable to provide consent; less than 16 years of age; or on interacting medications (e.g., antacids, cholestyramine, etc). Study subjects were asked to fast overnight before reporting to either the British Columbia Transplant Society Clinic or the Solid Organ Transplant Clinic on the study day. There were no restrictions on activity or food intake during the study day.  33  Following UBC Clinical Research Ethics Board approval (Appendix 1), written  informed  consent  (Appendix  2)  was  obtained.  Upon  administration of a steady-state morning mycophenolate mofetil (MMF) dose, an indwelling catheter was placed into a forearm vein of study subjects. Serial blood samples were then collected at 0, 0.33, 0.66, 1, 1.5, 2, 3, 4, 6, 8, 10, and 12 hours. The samples were collected in 3-mL tubes with ethylenediaminetetraacetic acid anticoagulant (BD Vacutainer, K3EDTA. BD – Canada, Oakville, ON). Each blood sample was ~3-5 mL. After each sample collection, the collection tube was inverted several times and kept on ice until processing. The catheter line was then flushed with saline solution (U.S.P 0.9% NaCl, Abbott Laboratory bacteriostatic injection, St. Laurent, PQ). About 1 mL of ―blood‖ was collected in the same manner and discarded immediately before the next sample collection to ensure no contamination with the saline solution. Plasma separation was achieved by centrifugation at 3000 rotations per minute for 5 minutes. Plasma was collected immediately into 2 empty vials, one of which was acidified (pH 2-4, 85% phosphoric acid solution; 10 µl per 500 µl of plasma) to preserve AcMPAG which is unstable at physiological pH. The samples were stored at -80ºC until analysis for MPA, MPAG and AcMPAG concentrations.  2.2.2 Plasma concentrations of MPA, MPAG , AcMPAG and fMPA The concentrations of MPA, MPAG, AcMPAG as well as fMPA were determined quantitatively in patient plasma samples by high performance liquid chromatography (HPLC) with ultraviolet detection. Full details of the methodology and method validation are provided in Appendix 3.  34  2.2.3 Assessment of pharmacokinetic parameters Conventional pharmacokinetic parameters, including AUC0-12, maximum concentration (Cmax), time to maximum concentration (Tmax), minimum concentration (Cmin), apparent oral clearance (Cl/F) and apparent volume of distribution (V/F), were calculated for each patient by noncompartmental analysis using the pharmacokinetic software WinNonlin Professional,  version  5.2  (Pharsight,  Mountain  View,  CA).  Dose-  normalized AUC as well as AUC ratios of MPAG/MPA and AcMPAG/MPA were also calculated.  2.2.4 Statistical analysis Descriptive statistics were used for demographic and pharmacokinetic data.  2.3 Results 2.3.1 Patient characteristics Sixteen stable adult islet transplant recipients (5 males, 11 females) receiving twice-daily MMF therapy were investigated.  Thirteen patients  were Caucasians, 1 was Japanese and 2 were of mixed ethnicities. Table 2-2 summarizes the patients’ characteristics. All patients have had diabetes for more than 5 years and had evidence of retinopathy and mild nephropathy at the time they received their islet-cell infusions. Patients were C-peptide negative at the time of transplant as well but became positive shortly after the transplant indicating that the transplanted islets were functioning. In addition to MMF, all patients were on a tacrolimus, steroid-free immunosuppressive regimen.  35  Table  2-1:  Characteristics  of  islet  transplant  patients  who  participated in this study CHARACTERISTIC  ALL SUBJECTS (N=16)  Gender  5 males, 11 females  Age (yrs)  50 ±8 (39-65)  Number of transplants per patient  2.7 ± 0.8 (1-4)  Insulin-free patients  9  C-peptide level (picomole/L)  423 ± 282 (47-1264)  HgA1C (%)  6.5 ± 0.7 (5.6-8.3)  Time since last transplant (days)  719 ± 720 (16-2006)  Height (cm)  166 ± 10 (150-183)  Weight (kg)  64 ± 11 (49-86)  MMF daily dosage (mg)  1609 ± 341 (1000-2000)  Serum creatinine (µmol/L)  100 ± 32 (51-170)  Albumin (g/L)  42 ± 3 (37-47)  Results expressed as mean ± standard deviation (range) HgA1c: hemoglobin A1c, normal range: 4.3% - 5.9%; C-peptide normal range: 270-1282 picomole/L  36  2.3.2 MPA Pharmacokinetics Table 2-2 summarizes the PK parameters of MPA, the metabolic ratios (MPAG/MPA and AcMPAG/MPA) and the free fraction of MPA in the study participants. Table 2-2: PK parameters and metabolic ratios of MPA in islet transplant recipients PHARMACOKINETIC  ALL SUBJECTS (N=16)  PARAMETER MMF dose (mg/kg/day)  25.4 ± 6.1 (15.9-39.2)  AUC (μg*h/ml)  42.9 ± 21.6 (11.3-88.6)  Dose-normalized AUC  52.9 ± 25.4 (15.1-118.1)  (μg*h/ml/g) Cmax (μg/ml)  13.0 ± 6.2 (2.9-22.4)  Tmax (hr)  1.2 ± 0.4 (0.7-2.0)  Cmin(μg/ml)  1.4 ± 1.0 (0.1-3.2)  Vd/F (L)  119.3 ± 76.2 (38.6-301.9)  Cl/F (L/hr)  25.0 ± 16.5 (8.5-66.2)  AUC ratio  17.8± 12.4 (6.3-43.5)  MPAG/MPA AUC ratio  0.1 ± 0.10 (0.0-0.3)  AcMPAG/MPA  37  PHARMACOKINETIC  ALL SUBJECTS (N=16)  PARAMETER Free fraction of MPA (%)  1.2 ± 1.0 (0.7-3.7)  Results expressed as mean ±standard deviation (range) AcMPAG: acyl glucuronide of MPA; AUC: area under the concentrationtime curve; Cl/F: apparent oral clearance; Cmax: maximum concentration; MMF: mycophenolate mofetil; MPA: mycophenolic acid; MPAG: 7-O-mycophenolic acid glucuronide;Cmax: maximum concentration observed in the 12-hour pharmacokinetic profile; Cmin: minimum concentration observed in the 12-hour pharmacokinetic profile; Tmax: time at maximum concentration; Vd/F : apparent volume of distribution  Three patients had undetectable levels of AcMPAG at all time points. Eight others had one or more undetectable AcMPAG levels throughout the 12hour sampling period. An ―undetectable level‖ was defined as one that is below the limit of HPLC detection of AcMPAG (i.e. 0.01 µg/mL) and was entered as a value of zero when calculating the pharmacokinetic parameters of AcMPAG, namely AUC0-12 (WinNonlin 5.2). The pharmacokinetic profiles for MPA, MPAG and AcMPAG are presented in Figures 2-3, 2-4, and 2-5. [Data points are expressed as mean ± SD]. Figures 2-6, 2-7 and 2-8 show the pharmacokinetic profiles of all 16 patients for MPA and its glucuronidated metabolites. There was large variability in MPA, MPAG and AcMPAG concentrations throughout the sampling period: 47.53%, 70.23% and 50.28% for Cmax, Cmin and AUC, respectively.  A re-absorption peak (for MPA) was  observed between 6-10 hours post-dose.  38  Figure 2-1: MPA pharmacokinetic profile of all 16 islet transplant recipients (mean ± SD)  20  MPA profile, islet transplant recipients (N=16)  18  MPA concentration (μg/ml)  16  14 12  10 8  6 4  2 0  0  2  4  6  8 time (hr)  39  10  12  Figure 2-2: MPAG pharmacokinetic profile of all 16 islet transplant recipients (mean ± SD)  40  Figure 2-3: AcMPAG pharmacokinetic profile of all 16 islet transplant recipients (mean ± SD)  AcMPAG profile, islet transplant recipients ( N=16)  5  AcMPAG concentration (μg/ml)  4  3  2  1  0  0  -1  2  4  6  Time (hr)  41  8  10  12  Figure 2-4: MPA pharmacokinetic profile, all 16 patients  IC01 IC02 IC03 IC04 IC05 IC06 IC07 IC08 IC09 IC10 IC11 IC12 IC13 IC14 IC15 IC16  MPA profile, islet transplant recipients (N=16)  24 22  MPA concentration (μg/ml)  20 18 16 14 12 10 8 6 4 2 0 0  2  4  6  8  time (hr)  42  10  12  Figure 2-5: MPAG pharmacokinetic profile, all 16 patients  180  IC01 IC02 IC03 IC05 IC06 IC07 IC08 IC09 IC10 IC11 IC12 IC13 IC14 IC15 IC16  MPAG profile, islet transplant recipients (N= 16)  MPAG concentration (μg/ml)  150  120  90  60  30  0 0  2  4  6  time (hr)  43  8  10  12  Figure 2-6: AcMPAG pharmacokinetic profile, all 16 patients  AcMPAG profile, islet transplant recipients (N=16)  6  AcMPAG concentration (μg/ml)  IC01 IC02 IC03 IC04 IC05 IC06 IC07 IC08 IC09 IC10 IC11 IC12 IC13 IC14 IC15 IC16  5  4  3  2  1  0 0  2  4  6  8  time (hr)  44  10  12  2.3.3 Free MPA The free MPA fraction is presented in Table 2-2. The mean free fraction was 1.18%, ranging from 0.73 to 3.74 %. There was no correlation between the free fraction and albumin levels (r= 0.01; data not shown).  2.4 Discussion 2.4.1 Inter-subject variability in pharmacokinetic parameters One of the essential aspects of successful organ/graft transplantation is effective immunosuppression. However, when the pharmacokinetics of an immunosuppressant are not predictable, the challenge for clinicians to prevent rejection on one hand and control adverse effects on the other can be great. Wide inter-patient variability in the pharmacokinetics of MPA has been reported in various transplant populations including the kidney, liver, heart and lung (78-81). Contrary to the liver and kidney, islet cells are not usually involved in drug elimination. Pharmacokinetics of MPA may thus be different in this subpopulation. In our study, we observed large interpatient variability in MPA PK as well as PK parameters for fMPA, MPAG and AcMPAG (Table 2-3, Figures 2-3 – 2-8). A brief comparison with recent studies done by various groups on the pharmacokinetics of MPA plus MPAG, AcMPAG and/or free MPA in various transplant populations is presented in Table 2-4(21, 65, 77, 81-84). Only one other study, conducted by Jacobson et al., was carried out in the islet transplant population(77). Because of the large number of studies recently published on the pharmacokinetics of MPA, selection criteria were narrowed down to those in transplant populations that have characterized, besides MPA, at least 2 of the following: MPAG, AcMPAG and/or free MPA. The comparison is meant to provide only an overview of 45  the pharmacokinetics of MPA and its metabolites in different transplant groups as patient demographics and study methods differ between centers. In general, inter-subject variability of MPA PK parameters was similar across the studies. The AUC, Cmax and Cmin in our patients were comparable to the other studies. The mean MPA AUC was within the recommended therapeutic range of 30-60 μg*h/ml (85). The MPAG/MPA AUC ratio varied from as low as 6.85 to as high as 63.7 between studies. Only a few studies included the analysis of AcMPAG(21, 81, 82). AcMPAG/MPA ratio was in the range of 0.076-0.417. Since the assays of MPA metabolites, especially AcMPAG, are a relatively recent development, information regarding expected metabolic ratios is yet in its infancy. The wide variability in the metabolic ratio of MPAG/MPA may reflect the interindividual variation not only in MPA metabolism but also in the enterohepatic recycling process in which MPAG conjugated in the liver and released into the bile is deconjugated in the intestine and reabsorbed as MPA(86).Plasma metabolic ratios of certain drugs are sometimes used as an indirect index of in vivo enzyme activity when these drugs are used as an in vivo probe(87). Whether MPA qualifies to be used as an in vivo probe wherein its metabolic ratios can be used to assess the in vivo activity of the UGT enzymes involved in its conjugation is an interesting question and may be the scope of future research studies. Unbound MPA fraction in our study was within the expected range and appeared to be in accord with other studies.  46  Table 2-3: Selected MPA pharmacokinetic studies from other research groups transplant  MMF  Concomi-  MPA AUC  Type  Dosage  tant  (μg*h/ml)  immuno-  MPA Cmax  MPA  MPAG/  AcMPAG/  (μg/ml)  Cmin  MPA  MPA  (μg/ml)  AUC ratio  AUC ratio  study  Islets  1-2 g  TAC  42.9±21.6  1.4±1.0  17.8±12.4  0.1 ± 0.1  1.2±1.0  (0.1-3.2)  (6.2-43.5)  (0.0-0.3)  (0.7-3.7)  Comments  fraction  suppressant  This  Free  13.0±6.1  of MPA  daily (11.3-88.6) (2.9-22.4)  47  N= 16  transplant  MMF  Concomi-  MPA AUC  Type  Dosage  tant  (μg*h/ml)  immuno-  MPA Cmax  MPA  MPAG/  AcMPAG/  (μg/ml)  Cmin  MPA  MPA  (μg/ml)  AUC ratio  AUC ratio  et al. (77)  Islets  0.75-1  SIR (with/  g BID  without TAC)  DAY 28  DAY 28  DAY28  NA  67 .8 (33.698.9)  Comments  fraction  suppressant  Jacobson  Free  DAY 28  18.6 (10.2- 1.22 (1.0540.1) 30.7)  7.70#  )  DAY 42  DAY 42  DAY 42  DAY 42  62.1 (11.593.9) DAY 60  19.4 (3.1427.0)  2.90 (0.603.75)  6.85# ()  DAY 60  DAY 60  DAY 60  of MPA  DAY 28 1.05 (0.781.57) DAY 42 0.85 (0.753.04) DAY 60  post- transplant. N=7 on 28, 42, 60 & 90 days, N=8 on 180 days N= 6 on 270 &  0.98 (0.831.52) DAY 90  360 days  0.91 (0.891.02) DAY 180  P>0.05 for all  33.6 (22.059.9) DAY 90  8.73 (1.8415.9)  1.16 (0.582.14)  NA ( )  DAY 90  DAY 90  DAY 90  64.7 (39.3-138)  16.5 (10.121.4)  2.61 (1.406.49)  NA (((((  DAY 180  DAY 180  DAY 180  DAY 180  49.0 (25.193.4)  12.4 (2.2618.5)  1.79 (0.533.37)  NA  1.14 (0.811.50)  DAY 270  DAY 270  DAY 270  DAY 270  DAY 270  51.2 (30.787.9) DAY 360  10.1 (6.6426.6)  1.62 (1.374.29)  NA  DAY 360  DAY 360  DAY 360  0.85 (0.741.38) DAY 360  43.4 (33.072.6)  10.7 (5.06-  2.16 (1.52-  NA  1.19  14.8)  3.43)  (0.851.34)  48  Days refer to  values at all study times  transplant  MMF  Concomi-  MPA AUC  Type  Dosage  tant  (μg*h/ml)  immuno-  MPA Cmax  MPA  MPAG/  AcMPAG/  (μg/ml)  Cmin  MPA  MPA  (μg/ml)  AUC ratio  AUC ratio  Lung  of MPA  0.5-1.5  TAC or  27.85  7.98  0.71 (unde-  13.84  0.36  1.9 (0.7-  g BID  CSA  (3.39-  (0.64-  tected-3.64)  (2.41-  (undetecte  3.8)  (with/with  115.31)  37.11)  55.19)  d-12.33)  al. (81)  Comments  fraction  suppressant  Ting et  Free  N=27 Values are dosenormalized.  -out steroids) Ting et  Heart  al. (81)  0.25-  TAC or  79.00  18.64  1.91 (0.34-  9.96  0.21  3.3 (0.2-  1.5 g  CSA  (16.89-  (3.62-  8.40)  (0.92-  (0.06-  15.0)  BID  with/with-  218.73)  47.28)  28.64)  2.03)  N=23 Values are dosenormalized.  out steroids) Kuypers et al.  Kidney  1 g BID  TAC  58.8 (27-  NA  1.95 (0.35-  111)  12.6#  0.076#  0.86  N=33  7.69) Only data from  (82)  12 months posttransplant are presented here. Values are dosenormalized.  49  transplant  MMF  Concomi-  MPA AUC  Type  Dosage  tant  (μg*h/ml)  immuno-  MPA Cmax  MPA  MPAG/  AcMPAG/  (μg/ml)  Cmin  MPA  MPA  (μg/ml)  AUC ratio  AUC ratio  Kidney  et al.  (diabetic  (21)  n=13)  Morning  46.72  dose:  ±45.52  MPA  DM) 2.32 ±2.52  ±10.28  ±240 mg  of  N =24 (13 had 11.72  654  Comments  fraction  suppressant  Akhlaghi  Free  12.50  0.376  0.9  (6.25-  (0.188-  (0.63-  33.33) ¥  0.752)  1.37)¥ P>0.05 for all  CSA/TAC  values between  with  diabetic and non-  steroids  diabetic patients  (non-  659  35.24  11.12  diabetic  ±280  ±17.92  ±8.63  1.67 ±1.33  n=11)  12.5  0.417  0.95  (6.25-  (0.237-  (0.54-  25.00)¥  0.833)  1.52) ¥  32.48#  NA  3.5 ±2.0  mg  Jiao  Kidney  0.75 g  CSA with  BID  steroids  20.2 ±6.5  9.4 ±3.4  262.3± 120.2x10-3  et al.(65)  N=12 Patients were on day 10 posttransplant.  Jain et al.(83)  Liver  1 g BID  TAC  40.0±  10.6± 7.5  30.9  1.1± 1.4  63.7±83.6  NA  3.9± 1.6  N=8  (2.0-3.6) Patients were within 1 month post-transplant. Values are  50  transplant  MMF  Concomi-  MPA AUC  Type  Dosage  tant  (μg*h/ml)  immuno-  MPA Cmax  MPA  MPAG/  AcMPAG/  (μg/ml)  Cmin  MPA  MPA  (μg/ml)  AUC ratio  AUC ratio  Free  Comments  fraction  suppressant  of MPA  dosenormalized. Pisupati et  Liver  0.5-1 g  TAC  118± 57.6  36.7± 15.6  NA  8.0±3.3  NA  1.9± 1.0  N=10  BID Patients were ≤  al.(84)  6 weeks posttransplant.  Values reported as Mean± SD or median,( range) unless otherwise indicated; AcMPAG: acyl glucuronide of MPA; AUC: area under the concentration-time curve(0-12 hours); BID: twice daily; Cmax: maximum concentration; CSA: cyclosporine; DM: diabetes mellitus; MMF: mycophenolate mofetil; MPA: mycophenolic acid; MPAG: 7-O-mycophenolic acid glucuronide; NA: not available; TAC: tacrolimus; SD: standard deviation * median # calculated by dividing median MPAG AUC by MPA AUC ¥ inter-quartile range  51  2.4.2 Comparison with the Jacobson study Jacobson et al. investigated the pharmacokinetics of mycophenolic acid over a one year period in 8 Caucasian women undergoing islet transplantation(77). The study had a longitudinal design in which they assessed  the  same  patients  over  the  one-year  follow-up  period  (specifically on days 28, 42 and then at 2,3,6,9 and 12 months). Full MPA PK profiles were obtained on days 28 and 42 post-transplant. Thereafter, a published LSS(56) (using plasma samples drawn within the first 6 hours post-dose) was applied to obtain PK parameters. This LSS was not fully validated before being applied as the authors only ascertained that the AUC estimations were not substantially biased. However, no limits were set as to how much bias was allowed, nor was the precision of estimations measured. When compared to our suggested LSS and validated using our patients’ profiles (section 3.3.3, Chapter 3), this LSS had a bias and precision of 19.26 % and 22.30%, respectively. The authors reported a large intra- as well as inter-patient variability that was highest for Cmax. Intra-subject variability was reported to be 63.7%, 55.3% and 37.3% for Cmax, AUC and trough concentration, respectively while inter-subject variability was reported as 66.6%, 40.4% and 58.7% Cmax,  AUC  and  trough  concentration,  respectively.  The  trough  concentration was defined as the time 0 (pre-dose) measurement. *Clinically, Cmin values are equivalent to trough concentrations and used interchangeably. The exact definition of each, however, is slightly different; Cmin value is the lowest measured drug level during the sampling period:  it could be the first (pre-dose) sample concentration  but it could also be the last sample taken right before the next dose, or, theoretically, any other concentration that happens to be the lowest in that sampling period(88). 52  The authors suggested alterations in MMF dose (as 6 patients required dose reductions prior to day 60 due to adverse effects), enterohepatic recycling or poor compliance as potential reasons for this variability. Changes in drug metabolism were considered as a less likely reason as apparent clearance estimates were stable over time. Dose-normalized AUC values were also reported to have stayed constant (p>0.05), which further supports the suggested explanation. Exact values, however, were not published in the study. Another observation of this study was a strong correlation between individual unbound (free) and total MPA concentrations (r2=0.94). A direct comparison of our study results with those of the Jacobson et al.(77)  shows  relatively  similar  observations  of  large  interpatient  variability in MPA PK parameters. In our study, the highest inter-patient variability was reported for Cmin at 70.23% followed by AUC and Cmax at 50.28% and 47.53%, respectively. As mentioned above, Jacobson et al. obtained full plasma profiles twice (days 28 and 42) and accordingly, MPAG AUCs were calculated twice as there was no LSS for estimating MPAG AUC when full profiles were not obtained. At both times, Jacobson et. al calculated MPAG metabolic ratios (7.70 and 6.85 on days 28 and 42, respectively) which were lower than our estimation of MPAG metabolic ratio at 17.77. However, it is hard to tell whether this difference is in fact significant because of the large variability in the estimated ratio as well as the intra-group variability in both studies. The Jacobson study also reported a strong correlation between free and total MPA concentration as mentioned above. In our study, we measured free MPA concentration only once, rather than at each sampling time. Therefore, we could not assess the correlation between free and total MPA concentrations. AcMPAG was not measured in the Jacobson study whereas we reported an AcMPAG metabolic ratio of 0.08. When 53  comparing  the MPA PK results of both studies, one could observe that  MPA AUC tends to be higher in the early post-transplant period (i.e. on days 28 and 42) in the Jacobson study compared to both the later posttransplant period (i.e. on days 60, 90, 180, 270, 360) in the same study and our study results. The authors suggest that the decline in AUC is most likely due to MMF dose reductions. This explanation appears reasonable as the MPA AUC values in our study were lower (42.89 μg*h/ml) and so was our mean dose (776 mg/dose). In contrast, the Jacobson study’s MPA AUC values were (67.8 and 62.1 μg*h/ml on days 28 and 42, respectively) and mean doses were 958 and 821 mg/dose on days 28 and 42, respectively). The patient populations in the two studies were different: the Jacobson study had a homogeneous smaller sample in which all 8 patients were of the same gender and ethnicity and were studied at the same time points post-transplant.  All  patients  had  also  initially  achieved  insulin  independence. Five of them maintained insulin independence throughout the study while graft function was lost in 3 patients on days >177, 207, and 330, respectively. Our patient population was more diverse and; included males and females studied from as early as 16 days posttransplant to as late as 2006 days; some were insulin independent while others were using a lower dose of insulin (than pre-transplant) at the time of the study. Our richer and larger sample has the potential to provide more inference about the pharmacokinetics of MPA in the islet transplant population, typically heterogeneous in its demographics. Both studies had different immunosuppressant regimens as well. Besides MMF, the Jacobson study used a corticosteroid-free sirolimus based regimen with or without tacrolimus. Our study patients were on MMFtacrolimus steroid-free regimens. Although in vitro studies showed that tacrolimus inhibited UGT enzymes(39, 40), emerging study data suggest 54  that MPA pharmacokinetic behavior in patients receiving concomitant sirolimus therapy is comparable to those on tacrolimus(28). Our close findings to the pharmacokinetic results reported by Jacobson et al. come to further support this fact. Our study explored, on a bigger scale, the pharmacokinetics of MPA as well as its glucuronidated metabolites: MPAG and AcMPAG. AcMPAG metabolic ratio is being reported for the first time in islet transplantation and we are among few studies that report this ratio in different transplant types(21, 81, 82). Our findings should be helpful in describing the pharmacokinetics of MPA as well as establishing the expected metabolic ratios of MPAG and AcMPAG in the islet cell transplant population. Our reported AcMPAG metabolic ratio may also be helpful in future studies that investigate whether AcMPAG levels do in fact have a clinically significant association with reported adverse effects of MMF therapy.  2.4.3 Intra-subject variability As this study was not longitudinal, intra-subject variability of MPA PK parameters was not determined. The longitudinal design of the Jacobson study (77) on the other hand, allowed for measurement of intra-patient variability (as discussed above). Large intra-subject variability in MPA pharmacokinetics has been reported by other transplant groups as well(58, 89-91).Pawinski et al.(58) reported a 90% increase with time in the MPA AUC within the first 3 months after kidney transplantation. The authors also reported a significantly higher trough (pre-dose) level of MPAG in the first week after transplantation. Mourad et al. (89) reported similar findings of  a  significant increase in the MPA AUC value between the early posttransplant period and at 3 months. On the other hand, Sanquer et al.(92) reported that the mean plasma pre-dose concentrations of MPA were 55  significantly lower in patients who had been on MMF for 2 to 3 years compared with those who had been on it for few months. This contradictory finding of intra-individual MPA pharmacokinetics shown to decrease rather than increase over time was addressed by Cox and Ensom(93) who speculated that the discrepant results may be explained by the fact that in the Sanquer study, the patients did not serve as their own controls as well and sample size of both groups (i.e. early and late transplant periods) were small. Furthermore, Ensom et al.(91) found no difference in any MPA pharmacokinetic parameter over three sampling periods over the first 9 months post-transplant in heart or lung recipients. The authors, however, discuss how their study had a major limitation of inability to synchronize the actual versus the target sampling periods. This difference may have contributed to the lack of a statistically significant difference in any of the measured MPA pharmacokinetic parameters.  2.4.4 Concomitant immunosuppressants Several concomitant immunosuppressive agents have been shown to affect the pharmacokinetics of MPA. Although the exact mechanism by which each immunosuppressant exerts its effects is still not fully understood, several suggestions have been made. Patients taking MMF with cyclosporine experience a lower exposure to MPA than those being treated with MMF alone or MMF in combination with tacrolimus or sirolimus(23). In specific, trough levels of MPA are reported to be significantly higher and secondary MPA re-absorption peaks (due to enterohepatic circulation) are more pronounced in patients receiving tacrolimus  or  sirolimus  co-therapy  compared  to  patients  on  cyclosporine(23). Cyclosporine is believed to decrease MPA exposure via impairing biliary excretion of MPAG into the intestine; the hypnotized 56  mechanism is that cyclosporine mediates this impairment via inhibition of the multi-drug resistance-associated protein 2 (MRP2; ABCC2), an active biliary transporter that transports MPAG from the bile to the intestine. Inhibition of MRP2, impairs the biliary excretion of MPAG, which in turn reduces the recirculation of MPA in the intestine and leads to decreased MPA and increased MPAG plasma concentrations(40, 94-96). The tacrolimus-MPA drug interaction is much less clear than the cyclosporine-MPA interaction(97). Zucker et al.(39) studied the in vitro inhibitory effects of tacrolimus and cyclosporine on UGT enzymes in human liver and kidney and showed that tacrolimus is a much stronger inhibitor of UGT enzymes than cyclosporine, thus suggesting that tacrolimus may potentially hinder MPA metabolism and lead to increased MPA levels. This in vitro study was fuelled by a previous study(98) that compared MPA plasma concentrations in groups of cyclosporine- and tacrolimus- treated patients and observed a higher Cmin and AUC in the tacrolimus treated group. It has been subsequently demonstrated that cyclosporine inhibits the biliary excretion of MPAG as mentioned in above. Whether or not tacrolimus increases MPA exposure can probably not be decided until further controlled studies in healthy volunteers are conducted (97). Data evaluating the effect of sirolimus on MPA glucuronidation is scarce(99). One in vitro study showed no effect of sirolimus on MPA metabolite  formation.  The  lower  MPA  exposure  when  MMF  was  administered with cyclosporine compared to when sirolimus is coadministered was attributed to the influence of cyclosporine on the excretion of MPAG into the bile(99). Other immunosuppressants usually included in prescribing protocols with MPA are corticosteroids. Corticosteroids can induce the expression of UGT 57  enzymes which may increase MPA metabolism (23)(100). However, it is still controversial whether this inductive effect is clinically significant; Cattaneo at al.(75) compared 26 renal transplant patients who were on MMF  and  cyclosporine  and  underwent  corticosteroid  tapering  and  eventual withdrawal to a control group of 12 patients who had the same immunosuppressive regimen but without steroid withdrawal. MPA AUC progressively increased as the corticosteroid was being tapered and eventually withdrawn. On the other hand, Kuypers et al. (101) followed 100  renal  transplant  recipients  of  whom  26  patients  underwent  corticosteroid tapering and eventual withdrawal within a year and found no significant effect of corticosteroid withdrawal on MPA AUC or trough concentrations. With these contradictory results, it is reasonable to conclude that only a randomized, prospective study could assess the corticosteroid-MPA interaction(100). Islet transplant immunosuppressant regimens are usually steroid-free and do not include cyclosporine. Tacrolimus is typically used instead of cyclosporine because, despite the fact that it was shown to increase posttransplant diabetes in kidney transplant, it was superior to cyclosporine in improving graft survival and preventing acute rejection(42). In fact, all subjects in this study were on the same immunosuppressant regimen; steroid-free-tacrolimus-MMF  combination.  Concomitant  immunosup-  pressant effects on MPA PK parameters is thus more likely to be comparable in all 16 subjects.  2.4.5 Disease states Diabetes is known to have an effect on both the pharmacokinetics and pharmacodynamics of drugs. Diabetes can cause a delay in gastric emptying  and  thus  may  affect  absorption  and  bioavailability  of  administered medications. It can also affect the amount or activity of 58  metabolizing enzymes(21). One prospective study compared the 12-hour trough levels of MPA between diabetic and non-diabetic kidney transplant recipients and found a statistically significant difference between the 2 groups (non-diabetic patients had almost double the mean trough level of diabetics: 2.94 µg/ml vs. 1.24 µg/ml for the 2 groups, respectively)(76). A case-controlled study compared the concentration – time profiles for MPA, free MPA concentration (fMPA), MPAG and AcMPAG in diabetic vs. non-diabetic kidney recipients. While no significant difference was observed in the characteristics of MPA or fMPA or the metabolites between the 2 groups, there was an initial slower absorption rate(20). Islet transplant recipients are also unique in their disease state. That is, they are a heterogeneous combination of patients who have had type I diabetes for a variable number of years and are currently totally insulinfree and on no anti-diabetic medications, insulin-free but taking oral hypoglycemics, or still using insulin but at a much lower dose. Various levels of glycemic control may add to the interpatient variability observed in our study.  2.5 Summary To our knowledge, this is only the second clinical study on the pharmacokinetics of MPA and MPAG in islet transplant recipients and the first study to address AcMPAG in this transplant population. We have observed large inter-patient variability in MPA, MPAG and AcMPAG pharmacokinetics  in  islet  transplant  recipients.  Factors  such  as  concomitant medication, disease state, patient demographics and lifestyle as well as genetic polymorphisms in UGT and ABCC2 may all play a role in explaining and predicting this variability and should be the aim of future studies in this field.  59  Chapter 3: Limited Sampling Strategies of Mycophenolic Acid for Estimation of Area under the Concentration-Time Curve in Islet Transplant Recipients 3.1 Specific Aim #2 The purpose of this study was to establish reliable and clinically convenient limited sampling strategies for the prediction of MPA areaunder-the-curve (AUC) in islet transplant recipients.  3.2 Methods 3.2.1 Patient population and MPA concentrations Please refer to sections 2.2.1 and 2.2.2, Chapter 2 for descriptions of patient population, sample processing and assay for determination of MPA concentrations.  3.2.2 Pharmacokinetic parameters assessment The AUC of MPA was determined for each study subject by noncompartmental  analysis  using  WinNonlin  Professional  version  5.2  (Pharsight, Mountain View, CA).  3.2.3 Limited sampling strategy determination and validation Multiple regression analysis was used to determine the LSSs. The Bayesian  method  could  not  be  used  because  no  population  pharmacokinetic data for MPA are currently available for islet transplant recipients. Without the population data, the Bayesian approach would 60  give predictions with a high amount of uncertainty (section 1.5.2.1, Chapter 1). Multiple regression analysis was performed using JMP 6.0.0 statistical software (SAS Institute Inc., Cary, NC). The AUC was the dependent variable and the timed concentrations were the independent variables. Preset criteria for selecting limited sampling equations were an R2≥ 0.75 and a maximum of 3 concentrations. Preset criteria for acceptable predictive performance were average bias and precision of no more than ±15%. Stepwise modelling was applied to calculate all possible multiple regression combinations. Of all resulting equations, those that met the preset selection criteria were considered for further validation. More than one approach was explored and different sampling-time limits were set until LSSs with satisfactory predictive performance, while still considered clinically convenient, were obtained. 3.2.3.1 The 2-group approach- untransformed data The 2-group approach was the first approach used to develop the LSSs. (section 1.3.3.2.a, Chapter 1). The 16 collected patients’ profiles were randomly split into two groups. One group (N=10) was assigned as ―the index group‖ and used to establish the limited sampling strategy. The other ―test or validation group― (N=6) was used to validate the developed LSSs. Only concentrations taken at or before 2 hours after drug administration were considered at this point and a maximum of three concentrations were used. This procedure (i.e. randomized splitting) was done twice and then LSSs were developed and validated for each randomization. This step was carried out to ensure that the results obtained were, in fact, reproducible. The randomization was achieved by an online random sequence generator(102).  61  3.2.3.2 The 2-group approach, log-transformed data Upon validation, LSSs developed using the 2-group approach had poor predictive performance. The concentrations and AUCs were then logtransformed in order to normalize the data for more reliable prediction and the entire process of developing LSSs was repeated again. (103). 3.2.3.3 Jackknife approach Log transformation still did not result in LSSs with satisfactory predictive performance results. Thus, the jackknife approach was the next method to apply (section 1.4.3.2.b, Chapter 1). Only concentrations taken at or before 2 hours after drug administration were considered at first and a maximum of three concentrations were used. This approach did yield LSSs with much better bias and precision values. Still, none of these developed LSSs met the predictive performance of within ±15%. As such, the next step was to relax sampling time criteria from within the first 2 hours to within the first 4 hours post-dose. Plasma sampling up to 4 hours post- dose may be less convenient than up to 2 hours. However, it is still clinically feasible and much more convenient than obtaining a full concentration - time profile(68, 72, 104). This step resulted in a remarkable improvement in bias and precision values and yielded LSSs with acceptable predictive performance. 3.2.3.4 Validation of LSS In the 2-group approach (both with and without log transformation), the validation group (i.e. 6 patients’ profiles) was used to validate the developed LSSs. The AUC predicted by each LSS was compared to the actual observed AUC. Bias and precision of the LSSs were then determined according to guidelines set by Sheiner and Beal(71). Relative bias was measured by the percent mean prediction error (%ME), and 62  relative precision was measured by the percent mean absolute error (%MAE), (section 1.4.3.1, Chapter 1). Acceptable accuracy and precision were set to be ≤± 15%(55). In the jackknife approach, each LSS was developed using all 16 profiles. In the validation step, each LSS equation was generated again N number of times, where N = 16. Each time it was derived from N-1 patients and used to predict the AUC of the Nth patient. Thus, a slightly different equation was used to predict the AUC of each patient. The AUC predicted by each patient’s LSS was compared to the actual observed AUC and bias and precision were determined in the same way as in the 2-group approach. 3.2.3.5 Comparison with other LSSs The LSS that was deemed most appropriate in the previous step was further compared for its predictive performance to other published LSSs in the renal transplant population. All patients’ profiles were used for the validation. Percent bias and precision were calculated, as well as number of patients within 15% bias and precision.  3.3 Results 3.3.1 The 2-group approach 3.3.1.1 Study subjects characteristics Characteristics of the islet transplant recipients in the index and validation groups in the first (A) and second (B) randomized-splittings are summarized  in  Tables  3-1  and  3-2,  respectively.  In  both  randomizations, no parameter was significantly different between the two groups.  63  Table 3-1: Characteristics of islet transplant recipients in the index and validation groups in randomization A ALL SUBJECTS  INDEX GROUP  VALIDATION GROUP  (N=16)  (N=10)  (N=6)  Male (number)  5  4  1  Age (yrs)  50.2 ± 8.3  51.8 ± .3  47.5 ± 6.2  Time since last transplant (days)  719 ± 720  758 ± 721  653 ± 782  Number of transplants  2.7 ± 0.8  2.6 ± 0.8  2.8 ± 0.8  Height (cm)  166 ± 10  167± 12  165 ± 5  Weight (kg)  64.1 ± 11.3  65.3 ± 13.3  62.2 ± 7.7  MMF daily dosage (mg)  1609 ± 341  1750 ± 333  1375 ± 209  Data expressed as mean ± SD  64  Table 3-2: Characteristics of islet transplant recipients in the index and validation groups in randomization B ALL SUBJECTS  INDEX GROUP  VALIDATION GROUP  (N=16)  (N=10)  (N=6)  Male (number)  5  2  3  Age (yrs)  50.2 ± 8.3  54.4 ± 7.4  47.8 ± 9.6  Time since last transplant (days)  719 ± 720  453 ± 698  1032 ± 503  Number of transplants  2.7 ± 0.8  2.9 ± 0.7  2.2 ± 0.8  Height (cm)  166 ± 10  165 ± 8  169 ± 13  Weight (kg)  64.1 ± 11.3  62.3 ± 9.1  68.0 ± 16.4  MMF daily dosage (mg)  1609 ± 341  1550 ± 307  1850 ± 224  Data expressed as mean ± SD 3.3.1.2 Untransformed data 3.3.1.2.a LSSs using single concentration In both randomizations, the correlations between AUC and single concentrations were generally poor [coefficient of determination or R 2 ranging  from  0.0468  to  0.7471  and  from  0.220  to  0.726  in  randomizations A and B, respectively]. The only exceptions were C10 (R2=0.854) and C4 (R2=0.754) in randomization A. As these singleconcentration LSSs were beyond the 2 hours post-dose criteria, they were not considered for further validation.  65  3.3.1.2.b LSSs using two and three concentrations In randomization A, twelve 3-concentration LSSs and two 2-concentration LSSs had acceptable R2 (>0.75) and sampling times within the first 2 hours post-dose. However, none of the 2- or 3-concentration LSSs had acceptable R2 values in randomization B. This clearly reflected how irreproducible these equations were. Upon validation, none of the equations that qualified had acceptable bias and precision values. Selected examples of equations developed using this approach are presented in Table 3-3.  66  Table 3-3: Selected examples of LSSs developed using the 2-group approach with the untransformed data  RANDOMIZ A-TION  NUMBER OF  LSS EQUATION  % BIAS  % PRECISION  CONCENTRATIONS USED  μg*h/ml for AUC  (% ME)  (% MAE)  μg/ml  for Cx  NUMBER OF PROFILES WITH OBSERVED VALUES WITHIN 15% OF PREDICTED VALUES*  A  2  AUC=17.559 + 6.107C0 + 1.761C2  14.08  49.06  1  A  3  AUC=21.057 + 8.619C0 -8.763C0.33 + 4.565C2  20.26  34.05  3  B  2,3  No equation met the criteria  NA  NA  NA  * Out of 6 validation pharmacokinetic profiles tested AUC: area under the concentration-time curve; ME: mean prediction error; MAE: mean absolute error; Cx: plasma concentration at time x; NA: not applicable  67  3.3.1.3 Log-transformed data 3.3.1.3.a LSSs using single concentration Again, in both randomizations, the correlations between AUC and single concentrations were generally poor [R2 ranging from 0.106 to 0.710 and from 0.171 to 0.646 in randomizations A and B, respectively]. The only exceptions were C10 (R2=0.792) and C4 (R2=0.773) in randomization A. As these single-concentration LSSs were beyond the 2 hours post-dose criteria, they were not considered for further validation. 3.3.1.3.b LSSs using two and three concentrations In  each  randomization,  nine  3-concentration  LSSs  and  two  2-  concentration LSSs had acceptable R2 (≥0.75) and sampling times within the first 2 hours post-dose (i.e. a total of 22 equations). Upon validation, none of the equations that qualified had acceptable bias and precision values. Bias and precision calculations were based on the predicted AUC value rather than the predicted log AUC; i.e. the prediction error was calculated as the difference between the actual AUC and predicted AUC values and not the difference between the log values. This is because AUC, rather than log AUC, is the parameter of interest for the estimation process. Furthermore, using the log values for bias and precision calculation markedly reduces the error and can be misleading if the results are interpreted as is rather than being transformed back to unlogged values by taking the anti-log function. Selected examples of equations developed using this approach are presented in Table 3-4.  68  Table 3-4: Selected examples of LSSs developed using the 2-group approach with logtransformed data RANDOMIZA -TION  NUMBER OF CONCENTRA-TIONS USED  LSS EQUATION  % BIAS  μg*h/ml for AUC  (% ME)  μg/ml  % PRECISION (% MAE)  for Cx  NUMBER OF PROFILES WITH OBSERVED VALUES WITHIN 15% OF PREDICTED VALUES*  A  2  log AUC= 1.059 + 0.275logC0 + 0.484logC1  108.11  108.11  0  A  3  log AUC=1.026 + 0.196 log C0 + 0.432Log C1+ 0.154 Log C2  16.55  28.47  3  B  2  logAUC= 0.982+0.086 log C0.66+0.200 log C1.5  -39.14  46.77  0  B  3  log AUC=1.026 - 0.059 log C0.33 + 0.117 log C0.66+ 0.203 log C1.5  -29.18  38.53  1  * Out of 6 validation pharmacokinetic profiles tested AUC: area under the concentration-time curve; ME: error; Cx: plasma concentration at time x  69  mean prediction error; MAE: mean absolute  3.3.2 Jackknife approach 3.3.2.1 Study subjects characteristics In this approach, all 16 patients’ profiles were used to develop the LSSs and the resulting equations were validated using the jackknife method. Characteristics of all patients are presented in the first column of Tables 3-1 and 3-2. 3.3.2.2.a LSSs using single concentration Again, the correlations between AUC and single concentrations were generally poor  [R2  ranging from 0.141  to 0.689 except for  C4  (R2=0.848)]. Validation of the LSS using only C4 yielded an acceptable bias of -8.01%. However, precision was 23.76%. Highly preferred, conventional sampling times at 0 or 2 hours post-dose did not yield satisfactory correlation with AUC. The R2 value was 0.595 and 0.596 for the C0 and C2 LSS, respectively. 3.3.2.2.b LSSs using two and three concentrations Two 2-concentration LSSs and eight 3-concentration LSSs had acceptable R2 (>0.75) and sampling times within the first 2 hours and were considered further for validation. A summary of validation results of these LSSs is presented in Table 3-5.  70  Table 3-5: Predictive performance of 2- and 3- concentration LSSs (within the 1st 2 hours post-dose) developed using the jackknife method R2  AUC EQUATION μg*h/ml for AUC μg/ml  % BIAS (%ME)  % PRECISION (% MAE)  for Cx  AUC= 8.993+4.146C0+1.137C1+1.892C2 AUC=13.023+5.110C0+0.686C1.5+1.631C2 AUC=12.868+4.671C0+0.602C0.66+2.260C2 AUC=16.189+6.961C0-1.704C0.33+2.495C2 AUC=8.925+1.074C0.33+1.614C1.5+2.127C2 AUC=9.257+1.611C1+0.157C1.5+2.363C2  AUC=9.355+0.059C0.66+1.65C1+2.480C2  71  NUMBER OF PROFILES WITH OBSERVED VALUES WITHIN 15% OF PREDICTED VALUES*  0.824  -14.03  30.53  7  0.788  -14.58  33.70  5  0.783  -15.90  33.39  5  0.781  -13.88  35.29  3  0.758  -17.52  36.58  4  0.751  -16.79  37.36  5  0.750  6.10  41.13  2  R2  AUC EQUATION  % BIAS  μg*h/ml for AUC μg/ml  (%ME)  % PRECISION (% MAE)  for Cx  AUC=11.633+6.763C0-1.534C0.66+2.531C1 AUC=15.843+5.702C0+2.151C2 AUC=9.386+1.699C1+2.464C2  NUMBER OF PROFILES WITH OBSERVED VALUES WITHIN 15% OF PREDICTED VALUES*  0.750  -14.75  37.86  6  0.768  -17.38  35.07  5  0.750  -15.98  35.05  5  * Out of 16 validation pharmacokinetic profiles tested AUC: area under the concentration-time curve; ME: error; Cx: plasma concentration at time x  72  mean prediction error; MAE: mean absolute  3.3.2.3 LSSs using relaxed sampling time criteria (up to 4 hours post-dose) None of the LSSs developed using sampling times within the first 2 hours post-dose met the preset predictive performance criteria of within ±15% bias and precision. The next step was to relax sampling time criteria from within the first 2 hours to within the first 4 hours post-dose. Including the sampling times at 3 and 4 hours in the LSSs modelling resulted in 54 equations that had acceptable R2 values. In order to narrow down the selection out of this large number of possible equations, equations with R2 ≥ 0.90 and with sampling times that are more conventional (i.e. at 0, 1, 1.5, 2, 3, 4 hours) were selected for further evaluation. There were three 2-concentration and nine 3-concentration LSSs that met these narrower conditions. A summary of the predictive performance of these LSSs is presented in Table 3-6.  73  Table 3-6: Predictive performance of 2- and 3- concentration LSSs (within the 1st 4 hours post-dose) developed using the jackknife method AUC EQUATION  R2  μg*h/ml for AUC μg/ml  for Cx  % BIAS (% ME)  % PRECISION (% MAE)  NUMBER OF PROFILES WITH OBSERVED VALUES WITHIN 15% OF PREDICTED VALUES*  AUC=1.783+1.248C1+0.888C2+8.027C4  0.980  -3.09  9.53  12  AUC=2.778+1.413C1+0.963C3+7.511C4  0.973  -3.22  11.02  13  AUC=1.448+1.239C1+0.271C1.5+9.108C4  0.960  -1.90  11.46  12  AUC=1.410-0.259C0+1.443C1+9.622C4  0.957  -2.68  11.53  12  AUC=7.135+0.671C1.5+0.788C2+8.522C4  0.921  -6.50  18.21  7  AUC=7.740+0.937C1.5+0.868C3+7.660C4  0.921  -5.86  18.27  9  AUC=6.644+1.168C0+0.902C1.5+8.536C4  0.912  24.85  27.48  4  AUC=7.509+2.069C2-1.246C3+10.109C4  0.910  -7.16  20.61  6  74  AUC EQUATION  R2  μg*h/ml for AUC μg/ml  for Cx  % BIAS (% ME)  % PRECISION (% MAE)  NUMBER OF PROFILES WITH OBSERVED VALUES WITHIN 15% OF PREDICTED VALUES*  AUC=9.667+1.631C0+1.234C2+7.827C4  0.909  -8.39  22.76  9  AUC=1.547+1.417C1+9.448C4  0.957  -2.46  11.14  12  AUC=6.515+0.967C1.5+9.347C4  0.907  -4.01  16.76  9  AUC= 9.849+1.303C2 +9.028C4  0.901  -9.16  21.56  8  *Out of 16 validation pharmacokinetic profiles tested AUC: area under the concentration-time curve; ME: error; Cx: plasma concentration at time x  75  mean prediction error; MAE: mean absolute  3.3.3 Recommended LSSs The only approach that yielded LSSs with acceptable bias and precision was the jackknife approach with sampling times up to 4 hours post-dose. One 2-concentration (at C1 and C4) and four 3-concentration (at C1,C2,C4; C1,C3,C4;  C1,C1.5,C4 and C0,C1,C4) LSSs had bias and  precision of less than 15%. All five equations had similar R 2 values, yielded close predictive performance results and included sampling up to 4 hours post-dose. However, the 2-concentration LSS was considered superior in that it utilizes one less sample, meaning reduced cost and less discomfort for the patient. It is, therefore, the one we recommend.  3.3.4 Comparison with other studies Selected MPA LSSs established in the kidney transplant population from other research groups are presented in Table 3-7. The renal population was chosen because it is well established and the most thoroughly studied.  76  Table 3-7: Selected MPA LSSs established in kidney transplant population from other research groups Reference  Suggested LSS  Population Studied  Method Used MRA  Johnson et al.(56)  AUC=9.02+3.77C0+ 1.33C1+1.68C3+2.96 C6  Yeung et al.(57)  AUC=5.19+6.92C0+ 1.08C1+0.72C2  Le Guellec et al.(59)  AUC=0.58C20min+ 0.97C1+6.64C3+3.48  10 patients studied on days 2,5 & 28 posttransplant used to develop the model in 10 additional patients (independent of those used to test the model) 10 patients studied at 1 MRA week, 1 and 3 months, and about 1 year posttransplant; analyzed retrospectively 20 patients at least 6 MRA months post-transplant  Pawinski et al.(58)  AUC=7.75+6.49C0+ 0.76C0.5+2.43C2  21 patients for a total of 50 plasma profiles  MRA  77  R2, Bias & Precision R2=0.841 Bias & precision: : NR  Notes  R2=0.765 Bias & Precision: NR  R2=0.946 Precision: 13.6% Bias: model is reported as showing no bias but no value is reported  Equation has same sampling times for MPA as well as cyclosporine; jackknife method applied for model development & validation  R2=0.862 Bias: 6.1 ± 19.0 Precision: NR  Equations developed using repeated cross-validation for randomly chosen subsets: data set was repeatedly randomly divided into two groups of 25 each: training and testing groups for a total of 50 times.  Reference  Suggested LSS  Population Studied  Van Hest et al.(61)  AUC=7.182+4.607C0 +0.998C0.67+ 2.149C2  133 patients; 7 of whom were diabetic (on days 7 & 11 posttransplant); analyzed retrospectively  Cho et al.(60)  Concentrations at time 0,1 & 8 hours post-dose showed significantly positive relationship Selected sampling times: 40 min, 1 h 30, 3 h (for AUC at day 7*) 20 min, 1 h, 3 (for AUC at day 30*) 20 min, 1 h, 3 h (for AUC after more than 3 months*) *time post-transplant Corrected AUC06: AUC=AUC 0-6h + 4C6 + 2C0 AUC=AUC 0-6h + 3C6 + 3C0  10 Korean patients  Premaud et al.(105)  Fleming et al.(62)  Method Used MRA  R2, Bias & Precision R2=0.75,0.67 Bias (mg.h/L): -1.5 (5.7-2.7), 0.2 1.31.6) Precision (mg.h/L): 6.9 (2.9-9.2), 8.1 (6.5-9.4) Means (95%CI) in diabetic, non-diabetic groups, respectively  NA  NA  44 patients: 24 de novo Bayesian patients had 2 profiles (on days 7 and 30) and 20 stable patients had 1 profile at > three months  R2=0.802, 0.859 & 0.893, & Bias: -5.7%,-8.2%& +0.4 on days 7, 30 & more than 3 months post-transplant, respectively  patients in the Indian subcontinent; 34 profiles from 31 patients divided to 4 groups according to treatment and time post-transplant, 3 groups evaluated for LSSs  R2 ~0.99. Bias: -1.17 ,0.25, 0.78 for the 3 different groups evaluated  MRA  Precision: NR  78  Notes Evaluated whether a limited sampling strategy developed and validated for non-diabetic patients can also be used in diabetic patients; Index data set contained 60 non-diabetic renal transplant recipients; Test set included remaining non-diabetic (66) and the 7 diabetic renal transplant recipients. Correlated AUC with one timed concentration and picked ones with significant positive correlation  Developed a corrected AUC0-6 to predict AUC012  Reference  Suggested LSS  Population Studied  Method Used MRA  Zicheng et al.(63)  AUC=12.61+0.37C0.5 +0.49C1+3.224 +8.17C10  31 Chinese patients  Zhou et al.(64)  AUC=14.81+0.80C0.5 +1.56C2+4.80C4 AUC=11.29+0.51C0.5 +2.13C2+8.15C8  75 Chinese patients on MRA day 14 post-transplant; 50 patients’ profiles used for index group; 25 patients’ profiles used for validation group  Jiao et al.(65)  AUC=10.403+0.841C 2+1.105C3+0.447C4 (for total MPA) AUC=180.543+0.956 C2−0.223C3+4.342C 4 (for fMPA)  12 Chinese patients on day 10 post-transplant  MRA  Muller et al.(66)  AUC=15.19+6.92C0 +1.08C1+0.72C2  18 stable patients  Correlation between actual AUC0-6 & calculated AUC0-12  Miura et al.(67)  AUC=0.26C0+2.06C2 +3.82C4+20.38  50 Japanese patients each had two 12-hour plasma profiles over 24 hours.  Mohammadpour et al.(72)  AUC=14.46C10+15.5  19 patients in early post-transplant period  MRA  MRA  79  R2, Bias & Precision R2:0.92 Bias & Precision:NR R2 = 0.70 for 1st equation R2 = 0.88 for 2nd equation Bias: of ±10.1 and ±6.9 mg. h/L for 1st & 2nd equations, respectively. R2=0.901 , 0.975 Bias:0.56±28.21,4.34 ±3.56 Precision: 11.22 ±0.94, 12.67±0.72 for equations 1 & 2, respectively R2=0.9082 Bias & Precision: NR  R2 = 0.693 Bias: 2.9% Accuracy: 17.1% Precision: 21.5%  R2=0.882 Bias: 0.17 mg.h/L Precision: 8.06 mg.h/L  Notes  Equations developed to simultaneously test MPA & fMPA; Jackknife method applied for model development & validation Algorithms taken from literature and correlated to AUC0-6 calculated to select the one with highest predictive power; selected equation was one developed by Yeung et al(42). Equation has same sampling times for MPA as well as tacrolimus; Profiles split randomly and evenly into validation & test groups; Each group had 50 profiles: 25 day & 25 night. Validation done by jackknife method  Reference  Suggested LSS  Figurski et al.(104)  AUC=8.31+5.91C0+ 0.79C0.66+5.86C4 (for patients on sirolimus)  Population Studied  Method Used 24 patients on sirolimus MRA 14 patients on cyclosporine  AUC=10.43+1.47C0+ 1.06C0.66+1.65C2 (for patients on cyclosporine)  R2, Bias & Precision R2=0.82 Precision=0.0 Predictive performance= 78% R2=0.86 Precision=4.1 Predictive performance= 83%  Notes Validation done by bootstrap method  AUC: area-under-the curve from 0-12 hours (AUC0-12); AUC0-6: area-under-the curve from 0-6 hours; CI: confidence interval; fMPA: free mycophenolic acid; LSS: limited sampling strategy; MPA; mycophenolic acid; MRA: multiple regression analysis; NR: not reported; NA: not applicable; R2: coefficient of determination  80  The predictive performance of the LSS deemed most appropriate (i.e. AUC=1.547+1.417C1+9.448C4) was further compared with the predictive performance  of  published  LSSs  derived  from  the  renal  transplant  populations. The comparison was done using all 16 patients’ profiles. Table 3-8 summarizes the comparison results. None of the published LSSs met our preset criteria for bias and precision (i.e. within ±15%). As expected when comparing the predictive performance of an LSS, generated using data from a particular subpopulation, to LSSs from the literature using the same set of data, our suggested LSS produced less biased and more precise predictions and had more patients’ profiles within the 15% limit of bias and precision (compared with the other published LSSs). Of all compared LSSs, only one (by Van Hest et al.(14)) was less biased than the suggested LSS. However, it was also less precise and had fewer profiles within 15% bias and precision.  81  Table 3-8: Predictive performance of previously published LSSs and results of comparisons in predictive performance between the recommended LSS for islet transplant patients and LSS derived from renal transplant populations Study  Population  Our equation Johnson et al.(56) Yeung et al.(57) Le Guellec et al.(59) Pawinski et al.(58) Van Hest et al.(61) Zicheng et al.(63) Zhou et al.(64)  Islet Renal Renal Renal Renal Renal Renal Renal  Jiao et al.(65) Renal Muller et al.(66)  Renal  Miura et al.(67)  Renal  Mohammadpour et al.(72)  Renal  AUC Equation  % pe (range)  % Bias (ME)  % Precision (MAE)  (-40.36-24.58) (-8.42-104.97)  -2.46 19.26  11.14 22.30  number of profiles with observed values within 15% of predicted values* 12 10  (-25.90-134.71)  25.99  34.85  7  AUC=0.58C20min+0.97C1+6.64C3+ 3.48 AUC=7.75+6.49 C0+ 0.76 C0.5+2.43C2 AUC=7.182+4.607C0+0.998C0.67+ 2.149C2 AUC=12.61+0.37C0.5+0.49C1+3.22C4 + 8.17C10 AUC=14.81+0.80C0.5+1.56C2+4.80C4  (-37.55-67.17)  2.50  25.31  6  (-35.01-47.29)  -2.83  21.13  7  (-40.02-55.39)  -0.77  21.01  6  (-36.06-61.18)  6.58  24.36  5  (-30.96-89.18)  8.08  31.82  3  AUC=11.29+0.51C0.5+2.13C2+8.15C8  (-35.85-42.13)  -2.71  17.14  9  AUC=10.403+0.841C2+1.105C3+0.44 7C4 AUC=15.19+6.92C0+1.08C1+0.72C2  (-62.28-23.02)  -41.47  45.77  1  (-25.90-143.71)  25.99  34.85  7  AUC=0.26C0+2.06C2+3.82C4+20.38  (-19.40-141.29)  20.63  32.17  7  (-42.31-15.57)  5.65  24.50  5  AUC=1.55+1.42C1+9.45C4 AUC=9.02+3.77C0+1.33C1+1.68C3+ 2.96C6 AUC=15.19+6.92C0+1.08C1+0.72C2  AUC=14.46C10+15.547  82  Study  Figurski et al.(104)  Population  AUC Equation  % pe (range)  % Bias (ME)  % Precision (MAE)  (-34.40-73.85)  10.68  24.33  Renal  AUC=8.31+5.91C0+0.79C0.66+5.86C4 For patients on sirolimus  number of profiles with observed values within 15% of predicted values* 6  (-54.24 - -53.39)  4.96  26.64  4  Renal  AUC=10.43+1.47C0+1.06C0.66+1.65C 2 For patients on cyclosporine  *Out of 16 validation pharmacokinetic profiles tested AUC: area under the concentration-time curve; ME: concentration at time x; Pe: Prediction error  mean prediction error; MAE: mean absolute error; Cx: plasma  83  3.4 Discussion 3.4.1 Current status of MPA LSSs The role of therapeutic drug monitoring (TDM) of MPA is still under investigation and evidence supporting TDM is still inconclusive (section 1.2.9, Chapter 1). Although controversial, it is generally recognized that MPA AUC is the most reliable index for acute rejection in solid organ transplant(31, 86, 106-109). Obtaining full AUCs can be clinically challenging for routine practice as it involves multiple sampling over a dosing period (usually 12 hours in the case of MPA). MPA pre-dose level (i.e. trough or C0), on the other hand, correlates poorly with AUC (31, 57, 79, 93). Therefore, LSSs for the estimation of MPA AUC in patients receiving MMF may potentially resolve the problem. In order for an LSS to be successful it has to be practical and clinically convenient (i.e. easy to implement, encourage patient adherence and involve minimum cost). It also has to provide estimates of AUC that are acceptably accurate and precise. LSSs of MPA AUC have been suggested by different groups(54, 56-67, 72, 74, 104, 105). Most of the published LSSs were developed in the renal and some in the heart, lung and liver transplant population. However, to our knowledge, no LSSs have been developed in the islet transplant population. The kidney, but not the islet cell, is involved in the metabolism of MPA. That could have a significant effect on the pharmacokinetics of MPA in these two populations. Furthermore, each LSS was center specific and developed in a different set of patients. Centers can have different immunosuppressant protocols, sampling times as well as analytical techniques to measure MPA. The enzyme multiplied immunoassay technique (EMIT), for example, overestimation of MPA AUC due 84  typically results in  to cross-reactivity with AcMPAG  compared to the HPLC technique(59). Each study recruits patients who may  be  of  various  transplant  types,  ages,  numbers,  ethnicities,  immunosuppressant regimens and co-morbid conditions. This versatility is reflected by different equations developed by each group using different sampling times. Also worth mentioning is the fact that some studies either did not report the predictive performance of their developed models(56, 57, 63, 66) and were satisfied with reporting the R or R2 values only or they partly tested the predictive performance of their LSS by reporting either bias or precision of their equations (but not both)(62, 64). The correlation coefficient is merely a demonstration of association between variables (i.e. AUC and timed concentrations) but it provides no information as to the accuracy or precision of the equation.(section 1.3.3, Chapter 1). Therefore, applying an LSS that was developed relying merely on an R2 value without full validation should be carefully considered. In our study, we developed LSSs for estimating MPA AUC in islet transplant recipients. The LSSs were developed and fully validated in this subpopulation providing a practical and clinically feasible tool to estimate MPA AUC precisely and accurately.  3.4.2 Approach used In this study, we aimed at developing LSSs to estimate MPA AUC in stable islet transplant recipients. The first approach we used was the 2group approach (with and without log transformation). Although very commonly used and easy to apply, the LSSs were not reproducible and more importantly, none of them had acceptable bias and precision values. That is likely due to the fact that our sample size was small and pharmacokinetics were highly variable from one patient to the other.  85  Using the jackknife approach yielded much better results because it utilized the data to its full richness. Although less widely used and much more time consuming, the jackknife method makes use of every patient’s profile in creating the LSS; then again each patient’s profile serves as a test in the validation step because it tests an equation that was developed using all patients’ profiles except for his/hers. Relaxing the sampling time to 4 instead of just 2 hours post-dose was another important factor in developing a successful LSS. MPA is absorbed mainly in the first 2 hours post-dose. Sampling times after 2 hours are likely to include more information about the elimination phase of MPA and give a better prediction of how large or small the AUC is expected to be. Although less convenient than LSSs developed with sampling up to 2 hours post-dose, LSSs utilizing sampling times up to 4 hours or more are still considered clinically feasible and more convenient than full AUC determination(60, 104). The LSS we suggest utilizes only 2 concentrations at conventional sampling times (i.e. 1 and 4 hours) and can be easily adopted by nursing staff. Utilizing sampling schemes up to to 10 hours(72) and up to 4 concentrations (56) can limit the clinical application of LSSs.  3.5 Summary To our knowledge, these are the first convenient and clinically convenient LSSs for MPA AUC prediction developed specifically in the islet transplant population. A total of 5 LSSs met our criteria of acceptable predictive performance and had conventional sampling times. Four utilized 3 concentrations and one utilized 2 concentrations. The LSS that we recommend  is  the  one  utilizing  two  concentrations:  AUC=1.547+1.417C1+9.448C4. This equation is convenient, clinically 86  feasible and may be useful for clinicians in optimizing patient care. Although validated preliminarily via the jackknife approach, further validation in larger numbers of islet transplant patients at our institution is planned. Other islet transplant centers may also wish to validate our equation in their population.  87  Chapter 4: Overall Summary and Conclusion 4.1 Overall discussion and conclusion There were two specific aims in this study: to characterize the pharmacokinetics of MPA and its glucuronidated metabolites, MPAG and AcMPAG; and to develop limited sampling strategies for estimation of MPA AUC in islet transplant recipients. Sixteen stable islet transplant recipients who were on MMF therapy were recruited for this 12-hour study. Patients were either insulin-free, or using a reduced insulin dose compared with what they were using before their transplant(s). In addition to MMF, all patients were on a tacrolimus-based steroid-free immunosuppressive regimen. There  was  large  interpatient  variability  in  all  pharmacokinetic  parameters of MPA, MPAG and AcMPAG. Three patients had no detectable AcMPAG levels at any point during the 12-hour sampling period. In comparison to the only other pharmacokinetic study in the islet transplant population which assessed MPA pharmacokinetic parameters throughout a 12-month period(77), MPA exposure (as indicated by AUC), and free MPA fraction were similar. MPAG metabolic ratio was higher in our study. However, it is hard to tell whether this was actually a significant difference rather than merely a result of the wide interpatient variability. It is generally accepted that the AUC of MPA over a dose interval serves as a good surrogate of drug exposure(28). Obtaining full AUC 88  profiles is, however, clinically challenging and costly; on the other hand, trough levels correlate poorly with AUC (section 1.2.9, Chapter 1). Therefore, LSSs for the estimation of MPA AUC in patients receiving MMF may help resolve the problem. All 16 patients’ profiles were used to develop the LSSs and then these developed LSSs were validated using the jackknife approach. The developed LSSs had to  have an R 2≥ 0.75, a maximum of 3  concentrations, and sampling times within 4 hours post-dose. The validation step included calculating the bias and precision of each developed LSS. Criteria for acceptable predictive performance were bias and precision within ±15%. Four 3-concentration LSSs and one 2-concentration LSSs met the criteria of acceptable predictive performance and had conventional sampling times. The LSS that we recommend is the one utilizing two concentrations: AUC=1.547+1.417C1+9.448C4. When compared with other LSSs published in the renal transplant population, the suggested LSS produced on average less biased and more precise predictions and had more patients’ profiles within the 15% limit of bias and precision. In fact, none of the published LSSs met our preset criteria for bias and precision (i.e. within ±15%) which indicates that LSSs may be center-specific and thus should be validated  before  being  applied  in  other  centers  or  transplant  populations. Our suggested equation is convenient, and more clinically useful for our islet transplant population than other LSSs that have been developed for renal transplant patients. Other islet transplant centers may wish to validate our equation in their population or use our study template as a  89  guide to develop accurate and precise LSSs specific to their own patient population.  4.2 Strengths and limitations Our study is the first to characterize the pharmacokinetics of MPA and its two glucuronidated metabolites MPAG and AcMPAG as well as free MPA in the islet transplant population. Furthermore, our results provided a convenient and clinically feasible LSS for estimating the MPA AUC in this transplant population. The developed LSS equation can be directly incorporated by clinicians in optimizing patient care. Our study had a basic research aspect as well as a clinical aspect. We strived to work in harmony between the two aspects, trying to explore the data to its full richness on one hand, while considering the clinical utility of our findings. The major limitation of our study is the small sample size of islet transplant recipients. The small sample size precluded us from investigating potential factors that might be contributing to the pharmacokinetic variability such as age, gender, time since transplant, insulin requirement, UGT enzyme polymorphisms, ABCC2 transporter polymorphisms and drug-drug interactions (as all the patients were on the same immunosuppressive regimen). As well, the small sample size necessitated use of the jackknife, instead of the 2-group, method in developing and validating the LSS. Thus, further more robust validation in a larger group of future islet transplant recipients at our center is planned.  90  Another limitation was that all our patients were stable; that means they were probably on an optimal MMF dose as their transplanted islets were functioning and they were reporting minimal side effects as well. As such, our study population would not have included those who had their MMF withdrawn due to severe side effects. Also, as this was a one-time PK study on stable patients, we were unable to correlate PK parameters to clinical outcomes.  4.3 Current knowledge Islet transplantation is a new and promising area in the treatment of type I diabetes. The islet transplant population has many unique characteristics and studies in this group are still lacking(77). The results obtained from this study should help better understand the pharmacokinetic behavior of MPA, a major immunosuppressant in this special population, and thereby lead to better patient care. However, many aspects of MPA pharmacokinetics remain unclear. One aspect is MPA’s pharmacokinetic-pharmacodynamic (i.e. treatment outcomes) relationship. This relationship has been relatively well established in the kidney transplant population; an MPA AUC of 30-60 μg.h/mL and trough of 1.0 – 3.5 mg/L are recommended (78, 85). This relationship, on the other hand, still needs to be established in the islet transplant population(77). Another aspect is the controversy regarding the value of MPA TDM in transplant patients and whether the choice of LSS to be used has an effect. Applying an LSS to estimate MPA exposure is beneficial only if that will yield favorable clinical outcomes (i.e. result in a lower incidence of graft rejection or side effects). Currently the evidence is inconclusive as to whether a relationship exists between AUC (as predicted by LSS) and either rejection or adverse effect occurrence(30). 91  A 2008 evidence report(30) investigated the evidence regarding whether TDM of MPA results in a lower incidence of transplant rejections and adverse events versus no monitoring of MPA. The report reviewed 495 published articles and found an equal number of studies demonstrating positive versus no associations between monitoring using limited sampling strategies and rejection. As for adverse events, there were more studies showing a lack of association rather than an association. Adding to the complication is the fact that most of the study designs were observational or case series developed with the intention of studying the biological or pharmacological effects of MMF dosing or MMF in combination with another immunosuppressant. None of the studies were designed to compare the incidence of rejection or adverse events in groups of patients whose MMF doses were controlled using different sampling strategies. These studies typically provide insight into the types of sampling strategies to use in monitoring, but they do not actually indicate whether monitoring and dose adjustment would have an effect on outcomes. Only two head-to-head randomized controlled trial (RCT) comparisons of ―concentration-controlled‖ techniques vs. fixed dose MMF have been published. The first RCT(32) compared (over a period of 12 months) kidney transplant patients who were randomly assigned to either a fixed dose of MMF that could be clinically adjusted or a concentration controlled dose that was adjusted to achieve an MPA set target level. There were more treatment failures (i.e. acute rejection, loss of graft or death) in the fixed dose group than in the concentration-controlled group.  The concentration-controlled group had significantly lower  treatment failures and acute rejections but there was no significant difference in incidence of most adverse effects. The second, more recent RCT compared in a similar manner, 2 groups of kidney 92  transplant patients either on a fixed dose MMF or clinically-adjusted dose but reported dissimilar results(33); over a 12-month follow up period, the authors observed  no statistically significant difference in  the incidence of treatment failure (which was a composite of biopsyproven acute rejection, graft loss, death or MMF discontinuation). However, they also reported that MMF dose adjustments based on target MPA exposure was not successful, partly because physicians seemed non-adherent to required changes in dose increments and rather reluctant to implement substantial dose changes.  4.4 Future research Because the islet transplant population is relatively small, future studies should be multicenter to provide a larger, richer pool of islet transplant patients. With a bigger sample size, the inter-patient variability in MPA pharmacokinetics can be more closely examined. Factors such as gender, age, time since transplant, insulin requirement could be separately investigated to see if one or more of these factors contribute significantly (and to what extent) to MPA pharmacokinetic variability. In addition, with a sufficiently large sample, the association between  polymorphisms  in  UGT  and  ABCC2  genes  and  the  pharmacokinetics of MPA can be studied. Results from such studies would  provide  more  insight  as  to  the  reasons  behind  MPA  pharmacokinetic variability. The pharmacokinetic-pharmacodynamic relationship of MPA in islet transplant patients is another area of future research. Studies aiming at establishing the therapeutic target range of MPA in the islet transplant population would help clinicians in their routine follow-up and ultimately lead to optimal patient care.  93  The  ultimate  goal would be  to  individualize  immunosuppressive  regimens, even before treatment begins, for islet transplant recipients for optimal treatment response and minimal toxicity. The combination of  pharmacokinetic,  pharmacodynamic,  and  pharmacogenetic  considerations would allow development of a dosing algorithm that integrates all aspects of MMF therapy. All of these prospective research studies will greatly improve our knowledge of MMF therapy in islet transplantation and will ultimately benefit islet transplant recipients.  94  5. References 1. Scientific summary of the collaborative islet transplant registry (CITR) for 2008 (www.citregistry.org) [homepage on the Internet]. 2. Robertson RP. Islet transplantation as a treatment for diabetes-A work in progress. N Engl J Med. 2004;350(7):694-705. 3. Gaglia JL, Shapiro AMJ, Weir GC. Islet transplantation: Progress and challenge. Arch Med Res. 2005;36(3):273-80. 4. Najarian JS, Sutherland DE, Baumgartner D, Burke B, Rynasiewicz JJ, Matas AJ, et al. Total or near total pancreatectomy and islet autotransplantation for treatment of chronic pancreatitis. Ann Surg. 1980;192(4):526-42. 5. Shapiro AMJ, Lakey JRT, Ryan EA, Korbutt GS, Toth E, Warnock GL, et al. Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen. N Engl J Med. 2000;343(4):230-8. 6. Fiorina P, Shapiro AMJ, Ricordi C, Secchi A. The clinical impact of islet transplantation. American Journal of Transplantation. 2008;8(10):19907. 7. Clinical guidelines for pancreatic islet transplantation program. British Columbia Transplant Society. September 2007. 8. Lehmann R, Spinas GA, Moritz W, Weber M. Has time come for new goals in human islet transplantation? American Journal of Transplantation. 2008;8(6):1096-100. 9. Gruessner RWG, Sutherland DER, Gruessner AC. Survival after pancreas transplantation. JAMA. 2005;293(6):675. 10. Warnock GL, Meloche RM, Thompson D, Shapiro RJ, Fung M, Ao Z, et al. Improved human pancreatic islet isolation for a prospective cohort study of islet transplantation vs best medical therapy in type 1 diabetes mellitus. Archives of Surgery. 2005;140(8):735-44.  95  11. Warnock GL, Thompson DM, Meloche RM, Shapiro RJ, Ao Z, Keown P, et al. A multi-year analysis of islet transplantation compared with intensive medical therapy on progression of complications in type 1 diabetes. Transplantation. 2008;86(12):1762-6. 12. Naftanel MA, Harlan DM. Pancreatic islet transplantation. PLoS Medicine. 2004;1(3):198-201. 13. Sutherland DER, Gruessner RWG, Gruessner AC. Pancreas transplantation for treatment of diabetes mellitus. World J Surg. 2001;25(4):487-96. 14. Ginns LC, Cosimi AB, Morris transplantation. Blackwell Science; 1999.  PJ.  Immunosuppression  in  15. Srinivas TR, Kaplan B, Meier-Kriesche HU. Mycophenolate mofetil in solid-organ transplantation. eop. 2003;4(12):2325-45. 16. Srinivas TR, Kaplan B, Meier-Kriesche HU. Mycophenolate mofetil in solid-organ transplantation. . 2003;4(12):2325-45. 17. Allison AC, Eugui EM. Mycophenolate mofetil and its mechanisms of action. Immunopharmacology. 2000;47(2-3):85-118. 18. Lee WA, Gu L, Miksztal AR, Chu N, Leung K, Nelson PH. Bioavailability improvement of mycophenolic acid through amino ester derivatization. Pharm Res. 1990;7(2):161-6. 19. Cellcept ®(mycophenolate mofetil) drug monograph. Nutley, NJ: Roche Laboratories; May 2008. 20. Ransom JT. Mechanism of action of mycophenolate mofetil. Ther Drug Monit. 1995;17(6):681-4. 21. Akhlaghi F, Patel CG, Zuniga XP, Halilovic J, Preis IS, Gohh RY. Pharmacokinetics of mycophenolic acid and metabolites in diabetic kidney transplant recipients. Ther Drug Monit. 2006;28(1):95-101. 22. Bullingham R, Shah J, Goldblum R, Schiff M. Effects of food and antacid on the pharmacokinetics of single doses of mycophenolate mofetil in rheumatoid arthritis patients. Br J Clin Pharmacol. 1996;41:513-6.  96  23. Staatz CE, Tett SE. Clinical pharmacokinetics and pharmacodynamics of mycophenolate in solid organ transplant recipients. Clin Pharmacokinet. 2007;46(1):13-58. 24. Shipkova M, Strassburg CP, Braun F, Streit F, Gröne HJ, Armstrong VW, et al. Glucuronide and glucoside conjugation of mycophenolic acid by human liver, kidney and intestinal microsomes. Br J Pharmacol. 2001;132:1027-34. 25. Picard N, Ratanasavanh D, Premaud A, Marquet P. Identification of the UDP-glucuronosyltransferase (UGT) isoforms involved in mycophenolic acid phase II metabolism. Drug Metab Disposition. 2004:139-46. 26. Ettenger R, Bartosh S, Choi L, Zhu W, Niederberger W, Campestrini J, et al. Pharmacokinetics of enteric-coated mycophenolate sodium in stable pediatric renal transplant recipients. Pediatr Transplant. 2005;9(6):7807. 27. Holt DW. Monitoring 2002;39:173-83.  mycophenolic  acid.  Ann  Clin  Biochem.  28. Shaw LM, Korecka M, Venkataramanan R, Goldberg L, Bloom R, Brayman KL. Mycophenolic acid pharmacodynamics and pharmacokinetics provide a basis for rational monitoring strategies. American Journal of Transplantation. 2003;3(5):534-42. 29. Shipkova M, Armstrong VW, Wieland E, Niedmann PD, Schütz E, Brenner-Weiß G, et al. Identification of glucoside and carboxyl-linked glucuronide conjugates of mycophenolic acid in plasma of transplant recipients treated with mycophenolate mofetil. Br J Pharmacol. 1999;126(5):1075-82. 30. Oremus M, Zeidler J, Ensom MH, Matsuda-Abedini M, Balion C, Booker L, et al. Utility of monitoring mycophenolic acid in solid organ transplant patients. Evid Rep Technol Assess (Full Rep). 2008 Feb;164:1131. 31. Hale MD, Nicholls AJ, Bullingham RES, Hené R, Hoitsma A, Squifflet JP, et al. The pharmacokinetic-pharmacodynamic relationship for mycophenolate mofetil in renal transplantation. Clinical Pharmacology & Therapeutics. 1998;64(6):672-83.  97  32. Le Meur Y, Buchler M, Thierry A, Caillard S, Villemain F, Lavaud S, et al. Individualized mycophenolate mofetil dosing based on drug exposure significantly improves patient outcomes after renal transplantation. American Journal of Transplantation. 2007;7(11):2496-503. 33. van Gelder T, Silva HT, de Fijter JW, Budde K, Kuypers D, Tyden G, et al. Comparing mycophenolate mofetil regimens for de novo renal transplant recipients: The fixed-dose concentration-controlled trial. Transplantation. 2008;86(8):1043-51. 34. Meiser BM, Pfeiffer M, Schmidt D, Reichenspurner H, Ueberfuhr P, Paulus D, et al. Combination therapy with tacrolimus and mycophenolate mofetil following cardiac transplantation: Importance of mycophenolic acid therapeutic drug monitoring. Journal of Heart and Lung Transplantation. 1999;18(2):143-9. 35. Meiser BM, Pfeiffer M, Schmidt D, Ueberfuhr P, Reichenspurner H, Paulus D, et al. In: The efficacy of the combination of tacrolimus and mycophenolate mofetil for prevention of acute myocardial rejection is dependent on routine monitoring of mycophenolic acid trough acid levels. Transplantation Proceedings; Elsevier; 1999. p. 84-7. 36. Flechner SM, Feng J, Mastroianni B, Savas K, Arnovitz J, Moneim H, et al. The effect of 2-gram versus 1-gram concentration controlled mycophenolate mofetil on renal transplant outcomes using sirolimusbased calcineurin inhibitor drug-free immunosuppression. Transplantation. 2005;79(8):926-34. 37. Zenapax® (daclizumab) drug Laboratories; September, 2005.  monograph.  Nutley,NJ:  Roche  38. Prograf® (tacrolimus) drug monograph. Deerfield, IL 60015-2548: Astellas Pharma US, Inc.; February, 2009. 39. Zucker K, Tsaroucha A, Esquenazi V, Tzakis A, Miller J. Evidence that tacrolimus augments the bioavailability of mycophenolate mofetil through the inhibition of mycophenolic acid glucuronidation. Ther Drug Monit. 1999;21(1):35. 40. Barten MJ, Shipkova M, Bartsch P, Dhein S, Streit F, Tarnok A, et al. Mycophenolic acid interaction with cyclosporine and tacrolimus in vitro and in vivo: Evaluation of additive effects on rat blood lymphocyte function. Ther Drug Monit. 2005;27(2):123-31.  98  41. Maes BD, Vanrenterghem YFC. In: Cyclosporine: Advantages versus disadvantages vis-à-vis tacrolimus. Transplantation Proceedings; Elsevier; 2004. p. 40-9. 42. Webster AC, Woodroffe RC, Taylor RS, Chapman JR, Craig JC. Tacrolimus versus ciclosporin as primary immunosuppression for kidney transplant recipients: Meta-analysis and meta-regression of randomised trial data. Br Med J. 2005;331(7520):810-21. 43. Rapamune® (sirolimus) drug monograph. Montreal, QC: Wyeth Canada; October, 2008. 44. King CD, Rios GR, Green MD, Tephly glucuronosyltransferases. Curr Drug Metab. 2000;1:143-61.  TR.  UDP-  45. Guillemette C. Pharmacogenomics of human UDPglucuronosyltransferase enzymes. The Pharmacogenomics Journal. 2003;3(3):136-58. 46. Mackenzie PI. Identification of uridine diphosphate glucuronosyltransferases involved in the metabolism and clearance of mycophenolic acid. Ther Drug Monit. 2000;22(1):10-3. 47. Kiang TKL, Ensom MHH, Chang TKH. UDP-glucuronosyltransferases and clinical drug-drug interactions. Pharmacology and Therapeutics. 2005;106(1):97-132. 48. Miners JO, McKinnon RA, Mackenzie PI. Genetic polymorphisms of UDP-glucuronosyltransferases and their functional significance. Toxicology. 2002;181:453-6. 49. Burchell B. Genetic variation of human UDP-glucuronosyltransferase: Implications in disease and drug glucuronidation. American Journal of Pharmacogenomics. 2003;3(1):37-52. 50. Marsh S, McLeod HL. Pharmacogenetics of irinotecan toxicity. Pharmacogenomics. 2004;5(7):835-43. 51. Stone AN, Mackenzie PI, Galetin A, Houston JB, Miners JO. Isoform selectivity and kinetics of morphine 3-and 6-glucuronidation by human UDP-glucuronosyltransferases: Evidence for atypical glucuronidation kinetics by UGT2B7. Drug Metab Disposition. 2003;31(9):1086-9.  99  52. Lampe JW, Bigler J, Horner NK, Potter JD. UDPglucuronosyltransferase (UGT1A1*28 and UGT1A6*2) polymorphisms in Caucasians and Asians: Relationships to serum bilirubin concentrations. Pharmacogenetics. 1999 Jun;9(3):341-9. 53. Bhasker CR, Mckinnon W, Stone A, Lo A.C K,T., Ishizaki T, Miners JO. Genetic polymorphism of UDP-glucuronosyltransferase 2B7 (UGT2B7) at amino acid 268: Ethnic diversity of alleles and potential clinical significance. Pharmacogenetics. 2000 Nov;10(8):679-85. 54. Ting LSL, Partovi N, Levy RD, Riggs KW, Ensom MHH. Limited sampling strategy for predicting area under the concentration-time curve of mycophenolic acid in adult lung transplant recipients. Pharmacotherapy. 2006;26(9):1232-40. 55. Ting LSL, Villeneuve E, Ensom MHH. Beyond cyclosporine: A systematic review of limited sampling strategies for other immunosuppressants. Ther Drug Monit. 2006;28(3):419-30. 56. Johnson AG, Rigby RJ, Taylor PJ, Jones CE, Allen J, Franzen K, et al. The kinetics of mycophenolic acid and its glucuronide metabolite in adult kidney transplant recipients. Clinical Pharmacology & Therapeutics. 1999;66:492-500. 57. Yeung S, Tong KL, Tsang WK, Tang HL, Fung KS, Chan HWH, et al. In: Determination of mycophenolate area under the curve by limited sampling strategy. Transplantation Proceedings; Elsevier; 2001. p. 10523. 58. Pawinski T, Hale M, Korecka M, Fitzsimmons WE, Shaw LM. Limited sampling strategy for the estimation of mycophenolic acid area under the curve in adult renal transplant patients treated with concomitant tacrolimus. Clin Chem. 2002;48(9):1497-504. 59. Le Guellec C, Büchler M, Giraudeau B, Le Meur Y, Gakoué JE, Lebranchu Y, et al. Simultaneous estimation of cyclosporin and mycophenolic acid areas under the curve in stable renal transplant patients using a limited sampling strategy. Eur J Clin Pharmacol. 2002;57(11):805-11. 60. Cho EK, Han DJ, Kim SC, Burckart GJ, Venkataramanan R, Oh JM. Pharmacokinetic study of mycophenolic acid in Korean kidney transplant patients. Journal of Clinical Pharmacology. 2004;44(7):743-50.  100  61. van Hest RM, Mathôt RAA, Vulto AG, Le Meur Y, van Gelder T. Mycophenolic acid in diabetic renal transplant recipients: Pharmacokinetics and application of a limited sampling strategy. Ther Drug Monit. 2004;26(6):620-5. 62. Fleming DH, Mathew BS, John GT, Chandy SJ, Manivannan J, Jeyaseelan V. A six-hour extrapolated sampling strategy for monitoring mycophenolic acid in renal transplant patients in the indian subcontinent. J Postgrad Med. 2006;52(4):248-52. 63. Zicheng Y, Peijun Z, Da X, Xianghui W, Hongzhuan C. Investigation on pharmacokinetics of mycophenolic acid in Chinese adult renal transplant patients. Br J Clin Pharmacol. 2006;62(4):446-52. 64. Zhou PJ, Yu ZC, Wang XH, Shao K, Zhao JP. Pharmacokinetics of mycophenolic acid and estimation of exposure using multiple linear regression equations in Chinese renal allograft recipients. Clin Pharmacokinet. 2007;46(5):389-401. 65. Jiao Z, Zhong J, Zhang M, Shi X, Yu Y, Lu W. Total and free mycophenolic acid and its 7-O-glucuronide metabolite in Chinese adult renal transplant patients: Pharmacokinetics and application of limited sampling strategies. Eur J Clin Pharmacol. 2007;63(1):27-37. 66. Muller H, Solari S, Zuniga C, Guerra I, Troncoso J, Ovalle R, et al. Therapeutic drug monitoring of mycophenolic acid in kidney transplant patients: A abbreviated sampling strategy. Transplant Proc. 2007 Apr;39(3):596-9. 67. Miura M, Satoh S, Niioka T, Kagaya H, Saito M, Hayakari M, et al. Limited sampling strategy for simultaneous estimation of the area under the concentration-time curve of tacrolimus and mycophenolic acid in adult renal transplant recipients. Ther Drug Monit. 2008;30(1):52-9. 68. Chen H, Gu Z, Chen B, Mao H, Zhang W, Fan Q. Models for the prediction of mycophenolic acid area under the curve using a limitedsampling strategy and an enzyme multiplied immunoassay technique in Chinese patients undergoing liver transplantation. Clin Ther. 2008;30(12):2387-401. 69. Ting LS, Partovi N, Levy RD, Ignaszewski AP, Ensom MH. Performance of limited sampling strategies for predicting mycophenolic acid area under the curve in thoracic transplant recipients. J Heart Lung Transplant. 2008 Mar;27(3):325-8. 101  70. Mahmood I, Miller R. Comparison of the Bayesian approach and a limited sampling model for the estimation of AUC and cmax: A computer simulation analysis. Int J Clin Pharmacol Ther. 1999;37(9):439-45. 71. Sheiner LB, Beal SL. Some suggestions for measuring predictive performance. Journal of Pharmacokinetics and Pharmacodynamics. 1981;9(4):503-12. 72. Mohammadpour AH, Nazemian F, Abtahi B, Naghibi M, Gholami K, Rezaee S, et al. In: Estimation of abbreviated mycophenolic acid area under the concentration-time curve during early posttransplant period by limited sampling strategy. Transplantation Proceedings; Elsevier; 2008. p. 3668-72. 73. Efron B, Tibshirani R. An introduction to the bootstrap. Chapman & Hall/CRC; 1993. 74. Kaczmarek I, Bigdeli AK, Vogeser M, Mueller T, Beiras-Fernandez A, Kaczmarek P, et al. Defining algorithms for efficient therapeutic drug monitoring of mycophenolate mofetil in heart transplant recipients. Ther Drug Monit. 2008;30(4):419-27. 75. Cattaneo D, Perico N, Gaspari F, Gotti E, Remuzzi G. Glucocorticoids interfere with mycophenolate mofetil bioavailability in kidney transplantation. Kidney Int. 2002;62(3):1060-7. 76. Zanker B, Sohr B, Eder M, Frohmann E, Land W. In: Comparison of MPA trough levels in patients with severe diabetes mellitus and from nondiabetics after transplantation. Transplantation Proceedings; Elsevier; 1999. p. 1167-. 77. Jacobson PA, Green KG, Hering BJ. Mycophenolate mofetil in islet cell transplant: Variable pharmacokinetics but good correlation between total and unbound concentrations. The Journal of Clinical Pharmacology. 2005;45(8):901-9. 78. Shaw LM, Holt DW, Oellerich M, Meiser B, van Gelder T. Current issues in therapeutic drug monitoring of mycophenolic acid: Report of a roundtable discussion. Ther Drug Monit. 2001;23(4):305-15. 79. van Gelder T, Klupp J, Barten MJ, Christians U, Morris RE. Comparison of the effects of tacrolimus and cyclosporine on the pharmacokinetics of mycophenolic acid. Ther Drug Monit. 2001;23(2):119-28.  102  80. Ting LSL, Partovi N, Levy RD, Riggs KW, Ensom MHH. Pharmacokinetics of mycophenolic acid and its glucuronidated metabolites in stable lung transplant recipients. Ann Pharmacother. 2006;40(9):1509-15. 81. Ting LSL, Partovi N, PharmD F, Levy RD, Riggs KW, Ensom MHH, et al. Pharmacokinetics of mycophenolic acid and its phenolic-glucuronide and acyl glucuronide metabolites in stable thoracic transplant recipients. Ther Drug Monit. 2008;30(3):282-91. 82. Kuypers DJ, Vanrenterghem Y, Squifflet JP, Mourad M, Abramowicz D, Oellerich M, et al. Twelve-month evaluation of the clinical pharmacokinetics of total and free mycophenolic acid and its glucuronide metabolites in renal allograft recipients on low dose tacrolimus in combination with mycophenolate mofetil. Ther Drug Monit. 2003;25(5):609-22. 83. Jain A, Venkataramanan R, Hamad IS, Zuckerman S, Zhang S, Lever J, et al. Pharmacokinetics of mycophenolic acid after mycophenolate mofetil administration in liver transplant patients treated with tacrolimus. The Journal of Clinical Pharmacology. 2001;41(3):268-76. 84. Pisupati J, Jain A, Burckart G, Hamad I, Zuckerman S, Fung J, et al. Intraindividual and interindividual variations in the pharmacokinetics of mycophenolic acid in liver transplant patients. The Journal of Clinical Pharmacology. 2005;45(1):34-41. 85. van Gelder T, Hilbrands LB, Vanrenterghem Y, Weimar W, de Fijter JW, Squifflet JP, et al. A randomized double-blind, multicenter plasma concentration controlled study of the safety and efficacy of oral mycophenolate mofetil for the prevention of acute rejection after kidney transplantation. Transplantation. 1999;68(2):261-6. 86. Oellerich M, Shipkova M, Schütz E, Wieland E, Weber L, Tönshoff B, et al. Pharmacokinetic and metabolic investigations of mycophenolic acid in pediatric patients after renal transplantation: Implications for therapeutic drug monitoring. Ther Drug Monit. 2000;22(1):20-6. 87. Tucker GT, Houston JB, Huang SM. Optimizing drug development: Strategies to assess drug metabolism/transporter interaction potential— toward a consensus. Pharm Res. 2001;18(8):1071-80. 88. Gibaldi M, Perrier D. Pharmacokinetics. Informa Healthcare; 1982.  103  89. Mourad M, Malaise J, Chaib Eddour D, De Meyer M, Konig J, Schepers R, et al. Correlation of mycophenolic acid pharmacokinetic parameters with side effects in kidney transplant patients treated with mycophenolate mofetil. Clin Chem. 2001;47(1):88-94. 90. Sanquer S, Breil M, Baron C, Dahmane D, Astier A, Lang P. Trough blood concentrations in long-term treatment with mycophenolate mofetil. Lancet(British edition). 1998;351(9115). 91. Ensom MHH, Partovi N, Decarie D, Ignaszewski AP, Fradet GJ, Levy RD. Mycophenolate pharmacokinetics in early period following lung or heart transplantation. Ann Pharmacother. 2003;37(12):1761-7. 92. Sanquer S, Breil M, Baron C, Dahmane D, Astier A, Lang P. Trough blood concentrations in long-term treatment with mycophenolate mofetil. Lancet (British edition). 1998;351(9115):640-8. 93. Cox VC, Ensom MHH. Mycophenolate mofetil for solid organ transplantation: Does the evidence support the need for clinical pharmacokinetic monitoring? Ther Drug Monit. 2003;25(2):137-57. 94. Westley IS, Brogan LR, Morris RG, Evans AM, Sallustio BC. Role of Mrp2 in the hepatic disposition of mycophenolic acid and its glucuronide metabolites: Effect of cyclosporine. Drug Metab Disposition. 2006;34(2):261-6. 95. Hesselink DA, van Hest RM, Mathot RAA, Bonthuis F, Weimar W, de Bruin RWF, et al. Cyclosporine interacts with mycophenolic acid by inhibiting the multidrug resistance-associated protein 2. American Journal of Transplantation. 2005;5(5):987-94. 96. Kobayashi M, Saitoh H, Kobayashi M, Tadano K, Takahashi Y, Hirano T. Cyclosporin A, but not tacrolimus, inhibits the biliary excretion of mycophenolic acid glucuronide possibly mediated by multidrug resistance-associated protein 2 in rats. J Pharmacol Exp Ther. 2004;309(3):1029-35. 97. Christians U, Jacobsen W, Benet LZ, Lampen A. Mechanisms of clinically relevant drug interactions associated with tacrolimus. Clin Pharmacokinet. 2002;41(11):813-51. 98. Zucker K, Rosen A, Tsaroucha A, de Faria L, Roth D, Ciancio G, et al. Unexpected augmentation of mycophenolic acid pharmacokinetics in renal transplant patients receiving tacrolimus and mycophenolate mofetil in 104  combination therapy, and analogous in vitro findings. Transpl Immunol. 1997;5(3):225-32. 99. Picard N, Premaud A, Rousseau A, Le Meur Y, Marquet P. A comparison of the effect of ciclosporin and sirolimus on the pharmokinetics of mycophenolate in renal transplant patients. Br J Clin Pharmacol. 2006;62(4):477-84. 100. Lam S, Partovi N, Ting LSL, Ensom MHH. Corticosteroid interactions with cyclosporine, tacrolimus, mycophenolate, and sirolimus: Fact or fiction? Ann Pharmacother. 2008;42(7):1037-47. 101. Kuypers DJ, Claes K, Evenepoel P, Maes B, Coosemans W, Pirenne J, et al. Long-term changes in mycophenolic acid exposure in combination with tacrolimus and corticosteroids are dose dependent and not reflected by trough plasma concentration: A prospective study in 100 de novo renal allograft recipients. The Journal of Clinical Pharmacology. 2003;43(8):866-80. 102. True random number service (www.random.org) [homepage on the Internet]. 103. Food US. Drug administration (FDA) guidance for industry: Statistical approaches to establishing bioequivalence. Washington (DC): FDA. 2001. 104. Figurski MJ, Nawrocki A, Pescovitz MD, Bouw R, Shaw LM. Development of a predictive limited sampling strategy for estimation of mycophenolic acid area under the concentration time curve in patients receiving concomitant sirolimus or cyclosporine. Ther Drug Monit. 2008;30(4):445-55. 105. Prémaud A, Le Meur Y, Debord J, Szelag JC, Rousseau A, Hoizey G, et al. Maximum A posteriori Bayesian estimation of mycophenolic acid pharmacokinetics in renal transplant recipients at different postgrafting periods. Ther Drug Monit. 2005;27(3):354-61. 106. Takahashi K, Ochiai T, Uchida K, Yasumura T, Ishibashi M, Suzuki S, et al. Pilot study of mycophenolate mofetil (RS-61443) in the prevention of acute rejection following renal transplantation in japanese patients. RS-61443 investigation committee-Japan. Transplant Proc. 1995 Feb;27(1):1421-4.  105  107. DeNofrio D, Loh E, Kao A, Korecka M, Pickering FW, Craig KA, et al. Mycophenolic acid concentrations are associated with cardiac allograft rejection. Journal of Heart and Lung Transplantation. 2000;19(11):10716. 108. Shaw LM, Korecka M, Aradhye S, Grossman R, Bayer L, Innes C, et al. Mycophenolic acid area under the curve values in African American and Caucasian renal transplant patients are comparable. The Journal of Clinical Pharmacology. 2000;40(6):624-33. 109. Nicholls AJ. Opportunities for therapeutic monitoring of mycophenolate mofetil dose in renal transplantation suggested by the pharmacokinetic/pharmacodynamic relationship for mycophenolic acid and suppression of rejection. Clin Biochem. 1998;31(5):329-33. 110. Shipkova M, Schutz E, Armstrong VW, Niedmann PD, Oellerich M, Wieland E. Determination of the acyl glucuronide metabolite of mycophenolic acid in human plasma by HPLC and emit. Clin Chem. 2000;46(3):365-72.  106  Appendices Appendix A: UBC Research Ethics Board's Certificate of Approval The University of British Columbia Office of Research Services Clinical Research Ethics Board – Room 210, 828 West 10th Avenue, Vancouver, BC V5Z 1L8 ETHICS CERTIFICATE OF EXPEDITED APPROVAL PRINCIPAL INVESTIGATOR:INSTITUTION / DEPARTMENT:UBC CREB NUMBER: Mary H. Ensom UBC/Pharmaceutical Sciences H06-03596 INSTITUTION(S) WHERE RESEARCH WILL BE CARRIED OUT: Institution Site: Vancouver Coastal Health (VCHRI/VCHA)Vancouver General Hospital Other locations where the research will be conducted: BC Transplant Society Clinic CO-INVESTIGATOR(S): Lillian Ting Garth Warnock Chantal Guillemette Rebecca Jean Shapiro Robert Mark Meloche Nilufar Partovi SPONSORING AGENCIES: Hoffmann-La Roche Ltd (Canada) PROJECT TITLE: PHARMACOKINETICS AND PHARMACOGENETICS OF MYCOPHENOLATE IN PANCREATIC ISLET TRANSPLANT RECIPIENTS THE CURRENT UBC CREB APPROVAL FOR THIS STUDY EXPIRES: December 21, 2007 The UBC Clinical Research Ethics Board Chair or Associate Chair, has reviewed the above described research project, including associated documentation noted below, and finds the research project acceptable on ethical grounds for research involving human subjects and hereby grants approval. DOCUMENTS INCLUDED IN THIS APPROVAL:APPROVAL DATE: Document Name Version Date Protocol: Clinical Research Protocol Islet Cell Tx MMF 1.0 November 7, 2006  107  Consent Forms: Consent Form Islet Cell Tx MMF Dec 18 20062.0December 18, 2006 Questionnaire, Questionnaire Cover Letter, Tests: PG MMF Islet cell Data Form Questionnaire 18Dec2006, MHHE N/A, December 18, 2006 Other Documents: Other: Please note that no one under the age of 19 yo will be recruited. As such, there will be no assent form [or signature line on the consent form for "guardian" (if subject <19yo)].December 21, 2006 CERTIFICATION: In respect of clinical trials: 1. The membership of this Research Ethics Board complies with the membership requirements for Research Ethics Boards defined in Division 5 of the Food and Drug Regulations. 2. The Research Ethics Board carries out its functions in a manner consistent with Good Clinical Practices. 3. This Research Ethics Board has reviewed and approved the clinical trial protocol and informed consent form for the trial which is to be conducted by the qualified investigator named above at the specified clinical trial site. This approval and the views of this Research Ethics Board have been documented in writing. The documentation included for the above-named project has been reviewed by the UBC CREB, and the research study, as presented in the documentation, was found to be acceptable on ethical grounds for research involving human subjects and was approved by the UBC CREB. Approval of the Clinical Research Ethics Board by: Dr. Gail Bellward, Chair  108  Appendix B: Consent form THE UNIVERSITY OF BRITISH COLUMBIA  Faculty of Pharmaceutical Sciences 2146 East Mall Vancouver, B.C. Canada V6T 1Z3 Tel: (604) 822-3183 Fax: (604) 822-3035 INFORMED CONSENT FORM PHARMACOKINETICS AND PHARMACOGENETICS OF MYCOPHENOLATE IN PANCREATIC ISLET TRANSPLANT RECIPIENTS           Principal Investigator Mary H.H. Ensom, Pharm.D., FASHP, FCCP, FCSHP, Faculty of Pharmaceutical Sciences, University of British Columbia and Department of Pharmacy, Children’s and Women’s Health Centre of British Columbia, (604) 875-2886 Co-Investigators R. Jean Shapiro, MD, FRCPC, Clinical Associate Professor, Department of Medicine, Faculty of Medicine, University of British Columbia; and Medical Manager, Solid Organ Transplantation, Vancouver General Hospital, (604) 875-5950 Nilufar Partovi, Pharm.D., FCSHP, Clinical Professor, Faculty of Pharmaceutical Sciences, University of British Columbia and Department of Pharmacy, Vancouver General Hospital, (604) 875-4293 Mai Al-Khatib, BSc(Pharm), MSc student, Faculty of Pharmaceutical Sciences, University of British Columbia, (604) 875-3198 Lillian Ting, MSc, PhD student, Faculty of Pharmaceutical Sciences, University of British Columbia, (604) 875-3198 Garth Warnock, MD, FRCSC, Professor and Head Department of Surgery, Faculty of Medicine, University of British Columbia, and Department of Surgery, Vancouver General Hospital (604) 875-4136 Mark Meloche, MD, FRCSC, Associate Professor, Department of Surgery, Faculty of Medicine, University of British Columbia and Surgical Transplantation, Vancouver General Hospital, (604) 875-5287 Chantal Guillemette, Ph.D., Faculty of Pharmacy, Laval University, (418) 656-4141 ext. 6348 Name and 24 Hour Telephone Number of Contact Person:Dr. Nilu Partovi (604) 8754293 Background You are being invited to participate in this study because you are a pancreatic islet transplant recipient and take the medication mycophenolate (Cellcept®). Mycophenolate is one of several immunosuppressive drugs used to help prevent rejection of your 109  transplanted pancreatic islet cells. You would be prescribed mycophenolate as standard treatment even if you did not participate in this study. Previous research studies have shown that different people handle mycophenolate in their bodies differently. Monitoring mycophenolate blood levels may help your doctor and pharmacist know what dose of mycophenolate works best for you. This means that knowing your mycophenolate blood levels may help determine whether there is enough drug in your body to prevent rejection but not too much to cause unwanted drug effects. Each person has a different genetic make-up. Therefore, one of the possible explanations for why different people handle mycophenolate differently may be genetic variation in the enzymes used to break down this drug in the body. Purpose The purpose of this study is to improve mycophenolate therapy in recipients of pancreatic islet cells by finding out how your body handles this drug. This study will determine if a person’s genetic make-up can explain why s/he may handle the drug differently than another individual. To achieve this, urine samples and 12 mycophenolate blood levels will be collected over a 12-hour period from 20 subjects. Study Procedures You will have the option of participating in this study if you are a pancreatic islet cell recipient, older than 19 years of age, who is being treated with mycophenolate (Cellcept®). (See also Exclusions below). If you choose to participate in the study, then the only procedure that will be different from usual transplant care is obtaining blood samples and total urine over a 12-hour period. You will be scheduled to visit the BC Transplant Society Office research clinic as an outpatient for your blood sampling. The study visit day will take approximately 13 hours. During the study visit day, you will be eating your usual breakfast, lunch, and dinner. Physical activity will be limited to walking within the building. A study visit appointment will be made with you by a study nurse or one of the investigators. You are asked not to eat anything after midnight on the evening before your morning appointment and report to the research clinic about half an hour before your usual morning dose of Cellcept®. When you arrive, the nurse will place a tiny catheter (identical to those used in the hospital after your transplant) into a vein in your forearm. This will allow easier blood collection and avoid having many “needle pokes” during the visit or you can choose to have “needle pokes” instead. If you have a “central line”, then all blood samples can be drawn from the “central line” instead of needing a catheter. You will have your first blood sample collected right before you take your usual morning dose of Cellcept®. After your usual morning dose of Cellcept®, you will have 11 more blood samples taken later at 20, 40, 60, 90 minutes, and at 2, 3, 4, 6, 8, 10, and 12 hours. All blood samples, except for one, require only about 3 ml (or one-half teaspoonful) each. For one time only, 110  an extra 20 ml (or 4 teaspoonsful) for the genetic analysis will be drawn during one of the blood sampling times. Thus, a total of about 56 ml (less than 2 ounces) of blood will be collected during the study visit day. In addition, you will be asked to collect all your urine during the 12-hour study period in provided containers. Your blood and DNA will not be used for any purpose except for those described in this current project. On the clinic visit day, a nurse will also complete a questionnaire with you to assess if you have had any unwanted drug effects with Cellcept®.  a) b) c) d)  Exclusions You must be excluded from study participation if: You refuse to or are unable to give written informed consent. You are younger than 19 years of age. Your mycophenolate (Cellcept®) therapy is not at steady state. “Steady state” means that you must have taken Cellcept® for at least 5 days without a dosage adjustment. You are taking other medications (e.g., antacids, cholestyramine, etc.) that can interact with mycophenolate. Risks The only risks associated with your participation in this study that are beyond your risks if you were not to participate would be the risks related to blood collection, repetitive and frequent sampling, and catheter placement. These risks are considered rare and mild but may include the following: slight bruising, temporary feeling of faintness, slight pain, collapsed vein and/or infection. There is also a hypothetical risk of disclosure of information related to your participation in this study. There may also be other adverse reactions or risks that could arise which are not predictable. If new information arises during your involvement in this study which could affect your desire to continue, you will be given such new information. See New Findings. Benefits The direct benefits to you as a participant of this study cannot be guaranteed, but may include improving your doctor’s understanding of how your body handles Cellcept® when your doctor, as a member of the investigative team, knows your study results. This information may be used to provide dosage recommendations specific for you. Your participation is also expected to help find whether genetic variation in the enzymes used to break down mycophenolate in the body can explain differences in handling of this drug by different people. Alternative Treatments If you decide not to participate or to withdraw at some later date, you will continue to receive Cellcept® as standard treatment to prevent rejection. Confidentiality Your confidentiality will be respected. No information that discloses your identity will be released or published without your specific consent to the disclosure. However, research 111  records and medical records identifying you may be inspected in the presence of the investigator or his or her designate by representatives of Health Canada and the UBC Research Ethics Board for the purpose of monitoring the research. However, no records which identify you by name or initials will be allowed to leave the investigators' offices. Remuneration/Compensation You will receive $100 total for the successful completion of this study to help offset costs for travel, meals, and parking required on the study visit day. There will be no other costs to you for participating in this study and you will not be charged for any research procedures. Compensation for Injury If you become ill or injured during the study, needed medical treatment will be available at no extra cost to you through your medical plan. Signing this consent form in no way limits your legal rights against the investigators, or anyone else. Contact If you have any questions, need more information about the study, or if you experience any adverse effects, you should contact Dr. Ensom at (604) 875-2886, Dr. Shapiro at (604) 875-5950, or Dr. Partovi at (604) 875-4293. If you have any concerns about your treatment or rights as a research subject, you may contact the Research Subject Information Line at the University of British Columbia Office of Research Services at (604) 822-8598. New Findings If you choose to enter this study and, at a later date, a more effective treatment becomes available, it will be offered to you. You will also be advised of any new information that becomes available that may affect your desire to remain in this study. Subject Consent Participation in this study is entirely voluntary and that you may refuse to participate or you may withdraw from the study at any time without any consequences to your continuing medical care. You do not have to provide any reasons for your decision not to participate or to withdraw. You have received a signed and dated copy of this consent form for your own records. You consent to participate in this study.  Subject Signature  Name (Print)  Date  Witness Signature  Name (Print)  Date  Investigator Signature  Name (Print)  Date  112  Appendix C: HPLC methodology and validation All work related to sample analysis including HPLC method development and validation was carried out at Dr. Mary Ensom’s laboratory at Children's and Women's Health Centre of British Columbia by lab. Technician Diane Decarie.  C.1 Plasma concentrations of MPA, MPAG , AcMPAG The  HPLC  instrumentation  (Waters  Alliance  System,  Waters  Ltd.,  Mississauga, ON) consisted of a delivery pump, an automatic injector equipped with a 200 µL injector loop, an Atlantis dC18 Symmetry C8 4.6 mm i.d. x 250 mm column, an Atlantis dC18 3.9 mm i.d. x 20 mm guard column and an ultraviolet detector set at 210 nm.  An integrator was  used to measure peak areas. MPA and metabolites concentrations were calculated as a ratio between their peak areas and that of the internal standard, indomethacin. Stock solutions of MPA, MPAG, AcMPAG, and indomethacin were prepared in HPLC-grade methanol. Calibration standards of the compounds were prepared by serial dilution in acidified plasma. Solutions of 1 mg/mL of MPA, AcMPAG, MPAG and internal standard (IS), indomethacin, were prepared in HPLC-grade methanol and kept at -20oC. Stock solutions of 100, 10, and 1 µg/mL of MPA and AcMPAG and 100 µg/mL of MPAG were prepared in HPLC-grade methanol and further diluted in extracted acidified serum to obtain the following standard concentrations:  0.25,  0.5, 1.0, 2.0, 5.0, 10.0, 15.0, 20.0, and 30.0 µg/mL for MPA; 0.2, 0.25, 0.5, 1.0, 2.0, 5.0, 10.0, 15.0, and 20.0 µg/mL for AcMPAG and 5.0, 10.0, 20.0, 30.0, 40.0, 60.0, 80.0, 90.0 and 100.0 µg/mL for MPAG.  Stock  solutions also contained 5 µg/mL of IS. All standards were prepared daily. The mobile phase consisted of a gradient of 0%-100%: 62%-38%: 0%100% (v/v) acetonitrile: 0.01M phosphate buffer (pH 3.0) at a flow rate 113  of 2 mL/min. All solvents and water were HPLC-grade and were filtered before injection of 75 µL of sample onto the column. The HPLC validation involved a calibration curve for MPA, AcMPAG and MPAG with 9 stock solutions and a blank at the beginning of each run. Calibration curves were generated by least-squares regression of the peak areas versus concentration of each stock solution.  The limit of  detection (LOD) for MPA was 0.02 µg/mL, for AcMPAG was 0.01 µg/mL, and for MPAG was 1.0 µg/mL. These were the lowest concentrations that gave integratable peaks. The lowest limit of quantitation (LLOQ) was set as the lowest concentration that could be quantified with acceptable precision and accuracy. Acceptable precision was set at an inter- and intra- coefficient of variation (CV) of ≤10% while acceptable accuracy was set at ≥80% when a minimum of four repeats were injected. Precision and reproducibility of the assay were evaluated by running the standards’ LLOQ, low, medium and high concentrations of 0.5, 4.0, 12.0, 25.0 µg/mL for MPA , 0.25, 3.0, 12.0, 18.0 µg/mL for AcMPAG and 10.0, 20.0, 50.0, 75.0 µg/mL for MPAG in quadruplicates daily for 4 days. Validation results are presented in Table 6-1. The assay’s inter- and intra- CV ranged from 0.8-10.4% for all compounds. Intra-day accuracy was 91.2, 90.0, 89.8 and 93.2% for MPA, 92.9, 94.7, 91.4 and 94.6% for AcMPAG, and 90.1, 89.6, 94.1, and 90.1% for MPAG at the LLOQ, low, medium, and high concentrations. Inter-day accuracy was 87.9, 99.4, 88.1 and 92.9% for MPA, 96.0, 97.03, 91.5 and 94.5% for AcMPAG, and 98.6, 84.0, 92.6 and 93.8% for MPAG at the LLOQ, low, medium, and high concentrations. Acceptable limits for the coefficients of variation were defined a priori as ≤ 10% and acceptable limits for accuracy as ≥ 80%.  114  Table C-1: Intra-day and inter-day coefficient of variation (CV) of MPA, AcMPAG and MPAG at four concentrations:  MPA  AcMPAG  MPAG  Concentration of standard used for validation (µg/mL)  Intra-day CV (%)*  Inter-day CV (%)*  0.5  2.5  7.8  4.0  2.6  6.0  12.0  5.4  6.8  25.0  2.1  2.7  0.25  5.3  9.2  3.0  2.2  8.3  12.0  1.9  2.5  18.0  2.1  4.7  10.0  6.4  10.4  20.0  0.8  2.5  50.0  3.3  3.1  75.0  3.5  7.2  *N= 4 AcMPAG: acyl glucuronide of mycophenolic acid; CV: coefficient of variation; MPA: mycophenolic acid; MPAG: mycophenolic acid glucuronide  Examples of calibration curves are provided in Figures C-1, C-2 and C3. While examples of chromatograms of calibration curve standards are provided in Figures C-4 and C-5. Samples were kept on ice for the duration of the extraction for total MPA, AcMPAG, and MPAG. Cold acetonitrile (1.2 mL at –20oC) containing 5 µg/mL internal standard was added to 300 μL of plasma sample and 115  vortex mixed. The supernatant was separated by centrifugation at 1250 g at 4oC and evaporated to dryness for 15 minutes at 37oC under 25 psi nitrogen flow. Samples were reconstituted in 300 μL of 20% acetonitrile. In addition, a 100 μL aliquot was further diluted 1:6 for detection of MPAG. Samples were filtered (Gelman 0.45 μm microfilter, Acrodisc GHP 13, Waters Ltd., Milford MA) before injection (50 μL) onto the HPLC column.  116  Figure C-1: Calibration curve of MPA R2 =0.99939 Calibration curve equation: Y= -4.27e-004x^2+1.12e+000X  40.00 30.00 Area Ratio 20.00 10.00 0.00 0.00  2.00  4.00  6.00  8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 Amount  117  Figure C-2: Calibration curve of MPAG R2 =0.998084 Calibration curve equation: Y=2.35e-003x^2+7.42e-001X  Area Ratio 40.00  20.00  0.00 0.00  5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00 50.00 55.00 60.00 65.00 70.00 75.00 80.00 Amount  118  Figure C-3: Calibration curve of AcMPAG R2 =0.996601 Calibration curve equation: Y= -8.22e-003x^2+9.47e-001X  15.00 Area Ratio 10.00 5.00 0.00 0.00  2.00  4.00  6.00  8.00  10.00 Amount  119  12.00  14.00  16.00  18.00  20.00  Figure C-4: HPLC chromatogram of a calibration curve standard containing MPA, AcMPAG and indomethacin: retention time of AcMPAG at 3.75 min , MPA at 4.85 min and indomethacin (IS) at 6.73 min  120  Figure C-5: HPLC chromatogram of a calibration curve standard containing MPAG; retention time at 3.39 min  121  C.2 Stability of AcMPAG AcMPAG is reported to be unstable under neutral and alkaline conditions (110). When plasma is acidified, the AcMPAG concentration has been shown to retain 90-100% of its original value for 24 hours at room temperature and for 30 days when kept at 4 and -20ºC (110). In this study, the stability of AcMPAG stored at -20, 4 and 25ºC was determined by analyzing an aliquot every week. AcMPAG was found to remain stable and maintained at least 88% of its original concentration for a period of up to 21 days. All plasma samples analyzed for AcMPAG concentration were therefore acidified (see section 2.1 in this Chapter).  C.3 Free MPA extraction The free fraction of MPA has been shown to be low (~3%) and the free MPA concentration (fMPA) is concentration-independent(23). Therefore, non-acidified plasma samples from subjects were pooled to obtain 1000 μL of plasma and the plasma was spiked with 25 μL of MPA stock solution (1 mg/mL) to ensure adequate free MPA concentrations within the analytical range of the HPLC assay. An aliquot of 400 μL of this spiked plasma was reserved for total MPA concentration quantification according to the procedure described above. For fMPA concentration, 500 μL of the spiked plasma was filtered with a Microcon YM-30 filter (30,000 molecular weight cut-off; Millipore, Billerica, MA) under centrifugation for 75 minutes at 4oC and 1380 g. An equal volume of 20% acetonitrile (in HPLC-grade water) containing 10 mg/mL indomethacin was added to the filtrate before injection (50 μL) onto the HPLC column. The free fraction was calculated by dividing fMPA by total MPA concentration in the spiked plasma.  122  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0067293/manifest

Comment

Related Items