* Abbreviations: AUC, area under the time/concentration curve; AUC 012, area under the time/concentration curve plotted using concentrations taken at 0, 1, and 2 hr; AUC 014, area under the time/concentration curve plotted using concentrations taken at 0, 1, and 4 hr; Cmax, maximal concentration of cyclosporine recorded during the pharmacokinetic study; GFR, glomerular filtration rate; 99mTc DTPA, technecium-99m diethylenetriamine pentaacetic acid; NEO, Neoral cyclosporine formulation; SIM, Sandimmune cyclosporine formulation; C0, blood concentration of cyclosporine immediately before the dose of SIM or NEO; Tmax, time at which Cmax occurred.
Renal transplantation of the diabetic patient, when compared with the nondiabetic patient, has been characterized by poor results, high patient mortality, and progression of the secondary complications of the diabetes. Many insulin-dependent diabetics come to end-stage renal failure with significant neuropathy, retinopathy, enteropathy, cystopathy, and cardiovascular disease. Selection for simultaneous pancreas and kidney transplantation has thus included careful evaluation of the life threatening complications. However, it has not precluded patients with the frequent, but less serious secondary complications, including gastroparesis and autonomic neuropathy, both of which may impact upon drug absorption.
Results of simultaneous pancreas and kidney transplants have shown good short-term and medium-term graft and patient survivals, although usually at the expense of increased use of immunosuppressant drugs(1). In a retrospective analysis of patients treated in this unit with simultaneous pancreas and kidney transplants, it was found that they have taken higher cyclosporine drug doses than age and sex matched renal recipients, but achieve comparable blood levels as measured by a parent-molecule specific radioimmunoassay. Given the known problems with gastroparesis in diabetic patients and the effect of metoclopramide on enhancing absorption of cyclosporine (2), a study was undertaken to examine the cyclosporine absorption profiles of stable renal pancreas transplant recipients using both the standard cyclosporine formulation Sandimmune (SIM*) and the recently available microemulsion formulation Neoral (NEO), both manufactured by Sandoz (Basel, Switzerland).
The aims of this study were first to determine whether there were changes in the absorption profiles using the 2 formulations and second to analyze the safety of conversion from SIM to NEO with respect to renal function and other side effects. The study was also designed to measure the dose conversion ratios achieved in stable simultaneous renal pancreas transplant recipients and examine simple monitoring schedules that could be used to assist in the management of such patients.
METHODS
Patients. Patients with stable and functioning simultaneous renal pancreas allografts more than 6 months after transplantation were eligible for entry. Seventeen patients, a mean of 4.75±2.0 years after transplantation, gave informed consent and were entered into the trial from which there were no withdrawals and 100% follow-up data. Patients were 45.2±7.4 years old and weighed 67.3±10.8 kg. All patients were free from exogenous insulin and had a serum creatinine of 151±33μmol/L at entry. All patients were taking azathioprine and prednisolone in addition to once daily SIM at recruitment. No patients were taking diltiazem, ketoconazole, or other drugs known to modify cyclosporine absorption or metabolism.
Study design. The study was approved by the Human Ethics Committee of the Western Sydney Area Health Authority and conducted to the standards of Good Clinical Practice. Sandimmune was converted from once to twice daily administration at week -4 by splitting the full daily dose equally to be taken at 8:00 a.m. and 8:00 p.m. each day. At week -2 pharmacokinetic, clinical, and renal function parameters were measured. Two weeks later, at week 0, SIM was replaced by NEO on a 1:1 conversion ratio. At weeks +2 and +4, the same pharmacokinetic, clinical, and renal function parameters were measured. Dosage adjustment at the start of the trial was based upon maintaining trough levels between 100 and 450 ng/ml or upon reduction in response to a rising serum creatinine value. After data acquisition from the initial patients, it became clear that 12-hr trough levels could fall below the planned target level despite absorption of a considerably increased total amount of drug. The criteria for dose increase were thus modified to include both a trough level below 100 ng/ml and decreased area under the time/concentration curve (AUC).
Drug formulations and monitoring. Sandimmune capsules (SIM)(Sandoz, Basel, Switzerland) were used as standard cyclosporine therapy before conversion. Cyclosporine in microemulsion formulation (Neoral, NEO) was provided by Sandoz, Basel, in both 100-mg and 25-mg strength capsules(3). All medication was dispensed by the hospital pharmacy, and both used and unused packs were returned to confirm compliance. Each patient kept a daily drug administration diary with a record both of drug doses and times of administration.
Blood levels of cyclosporine were measured in batches using a radioimmunoassay (CYCLO-Trac SP, whole blood sample, IncSTAR, Stillwater, MN). Trough levels were measured in the clinic before the morning dose of cyclosporine. An abbreviated time/concentration AUC pharmacokinetic profile was measured at weeks -2, +2, and +4 with blood samples taken before the morning dose (time 0) and then at 0.5, 1, 1.5, 2, 4, 6, and 9 hr after the dose. C0 was the blood concentration of cyclosporine immediately before the dose of SIM or NEO and thus also represented the trough value from the dose taken the previous evening; Cmax was the maximum concentration recorded during the pharmacokinetic study; Tmax was the time at which Cmax occurred. AUC was the area under the time/concentration curve to 9 hr, and AUC 012 and AUC 014 were the trapezoid areas plotted using concentrations in the samples taken at 0, 1, and 2 hr and 0, 1, and 4, hr, respectively.
Renal function and clinical monitoring. Serum creatinine was measured routinely using an autoanalyzer employing the Jaffe method (Hitachi 747 automatic analyzer, Tokyo, Japan, and Boehring-Mannheim reagents). Plasma technecium-99m diethylenetriamine pentaacetic acid (99mTc DTPA) clearance was measured from 2 timed blood samples taken after a single injection (4), and the glomerular filtration rate (GFR) was calculated and adjusted for body surface area(weight0.425×height0.725×0.007184). The GFR was measured during the first 4 hr after the administration of the cyclosporine formulations, and blood samples were taken simultaneously for the GFR and pharmacokinetic assays to reduce the number of sampling times for the patients.
Clinical history and examination were performed on each clinic visit and recorded. Dose adjustments were either made at the clinic visit or by telephone within 1 day of the visit depending upon when the serum creatinine and drug blood concentrations were available.
Statistical analyses. Data were entered onto an SPSS data file directly from the case report forms, and data accuracy was validated. Subsequent analyses were undertaken using the Statistical Package for Interactive Data Analysis (Statistical Computing Laboratory, Macquarie University, Eastwood, Australia). Because this was the first analysis of this formulation of cyclosporine in this type of patient, exploratory data analysis was used to examine all potentially relevant variables. A P value of<0.05 was used to determine the level of significance. Analyses of all pharmacokinetic parameters were performed using raw data, dose adjusted data, and dose per kilogram adjusted data. AUC was calculated using the trapezoid method. Comparison of absorbers was undertaken using McNemar's test, the paired t test examined changes in renal function and pharmacokinetics, and Spearman's rank correlation was used to examine associations between the changes of parameters before and after conversion from SIM to NEO. Linear regression analysis was used to analyze predictors of AUC.
RESULTS
Safety and tolerability. Conversion from SIM to NEO was achieved in all 17 patients according to protocol, and no patient has been withdrawn from the trial at any stage. Data collected at weeks -2, +2, +4, and +12 from all patients are shown in Table 1. With respect to the effect of conversion on allograft function, no patient used insulin at any stage during the trial period. Renal function, as measured both by serum creatinine and GFR, was not significantly altered for the group as a whole. There were however, individual patients with changes in the serum creatinine values and the GFR. The serum creatinine rose by more than 15% in 6 patients with consequent protocol driven downward adjustment in NEO dose. The GFR rose in 3 patients, did not change in 8, and fell in 6. Detailed analysis of the changes in GFR revealed a single significant correlation between changes in the pharmacokinetic parameters and GFR with a reduction in Tmax showing a negative correlation with a fall in GFR only at 2 weeks and using dose adjusted data (r=-0.50, P<0.04). There were no statistically significant correlations with changes in serum creatinine, drug dose, dose per kilogram, Cmax, AUC, or C0 at either 2 or 4 weeks.
No adverse events were attributed to the conversion to NEO; a total of 4 urinary tract infections, 4 upper respiratory tract infections, 4 patients with mild gastroenteritis, 1 superficial skin infection, and 1 foot ulcer occurred in the 34 patient months from week 0 to +8. There were no serious adverse events, though 1 patient was admitted to the hospital for care of the neuropathic foot ulcer.
Drug doses and pharmacokinetics. Whereas the conversion ratio established by the protocol was 1:1 from SIM to NEO, doses were reduced during the first 4 weeks of therapy with NEO and continued to be reduced slowly during the remaining 8 weeks, as shown in Table 1. In the group overall there was a 17% mean reduction in dose (18% mean reduction in dose per kilogram) by week +4, with an actual dose reduction in 7, increase in 2, and no change in 8. The 2 patients in whom the dose was increased were recruited in the early part of the study, and dose adjustment was made by protocol on the basis of a 12-hr trough level of <100 ng/ml. It was apparent in both patients, when the AUC values were available, that, although the trough levels had dropped slightly and to below the predetermined range, the AUC values had actually increased after conversion to NEO. To avoid still further increasing the patients' total exposure to cyclosporine, trough levels in the range 80-100 ng/ml were ignored in subsequent patients unless the AUC had also dropped. Because this did not occur, there were no further patients in whom the dose was increased on conversion to NEO. Dose reduction, on the other hand, was driven by a rise in the serum creatinine of >15%, because no patients had trough levels that exceeded the protocol range. Given the small number of patients, it was not surprising that no significant correlations were found between dose change and measured parameters of renal function. Figure 1 shows the raw data for the changes in AUC between weeks -2 (SIM) and +2 and +4 (NEO) for each individual patient.
Analysis of the individual patients' pharmacokinetics with SIM demonstrated that arbitrary criteria of Cmax>500 ng/ml and Tmax≤2 hr divided 5 patients with “good absorption” using the conventional formulation from 12 who did not meet both of those criteria. After conversion to NEO, 14 at week +2, and 13 at week +4 met both criteria for good absorption(P<0.02 and <0.01, respectively). There was no difference between the 5 good absorbers of SIM and the remaining patients in terms of their early transplant course with a mean of 2.6 and 2.8 rejection episodes per patient, respectively, and no significant difference in utilization of monoclonal antibody rejection therapy.
The pharmacokinetic profiles of the 2 formulations were significantly different with respect to most of the measured parameters.Figure 2 shows the mean percentage change in AUC, C0, and Cmax. There were no differences when the raw data were adjusted for drug dose or dose per kilogram, and the dose adjusted figures are thus shown. The changes in Tmax were quite variable with a median reduction in Tmax of 50% from 3.2±1.9 hr at week -2 with SIM to 1.8±1.3 hr at week+2 on NEO (P<0.007). The spread of Tmax values appeared to decrease after conversion to NEO (P<0.06), but there was an increase in the spread of Cmax values using dose adjusted data(P<0.03).
The pharmacokinetics of NEO were more predictable than those of SIM, although there was only a weak correlation between the dose administered and the AUC (Table 2). There were no correlations between dose or dose per kilogram and C0, Cmax, or Tmax with either formulation at any time point. Examination of the changes between the pharmacokinetic profile for each patient on SIM and NEO at each time point revealed a correlation between the increase in AUC and the increase in C0 (r=0.63 and P<0.01, and r=0.55 and P<0.05 for the changes between weeks -2 and +2, and -2 and +4, respectively).
The final analysis of the pharmacokinetics centered around the possibility of using a reduced series of blood concentration measurements rather than the 9-hr profiles used in this study. Renal pancreas transplant recipients absorbed cyclosporine from SIM less effectively, on the whole, than from NEO; however, there was considerable variability between patients, thus, it was important to determine whether there may be a simple method of examining that variability. As already shown, the actual change in trough concentrations was small but correlated weakly and significantly with change in AUC. Three sample points were selected as practical in the context of normal clinical practice, and AUCs were thus calculated for samples taken at 0, 1, and 2 hr to provide“AUC 012,” and at 0, 1, and 4 hr to provide “AUC 014.” The correlation between these reduced AUCs and the 9-hr profile showed that most of the variability in results could be captured by these limited profiles(Table 3). Whenever there was a significant correlation with AUC or change in AUC, there was also a significant correlation or change in both AUC 012 and AUC 014. The implication from this analysis is thus that either reduced profile would significantly correlate with the 9-hr profile, neither being clearly superior. There was, however, 1 important difference between these profiles and the full 9-hr AUC, as shown inFigure 2. The mean percentage change in AUC was 87-89%, whereas the change in AUC 012 was 265-280% and that in AUC 014 was 159-168%. In other words, most of the change in AUC was seen in the first 2 hr and almost all in the first 4 hr. The danger of the reduced profiles is that they could be misinterpreted as showing dramatically increased total exposure to the drug as a result of conversion from SIM to NEO, overestimating the AUC 012 by over 3-fold and the AUC 014 by almost 2-fold.
Despite exhaustive stepwise analysis of the preconversion factors that might predict the pharmacokinetic profiles of renal pancreas transplant patients converted to NEO from SIM, none of the fitted models were found to have an R2 value of greater than 0.4. The conclusion from this analysis was that there was no substitute for actual measurement of blood levels with subsequent dose adjustment when converting such patients.
DISCUSSION
This study has confirmed that previously diabetic recipients of simultaneous renal pancreas allografts have both very variable and poor absorption of cyclosporine when using the conventional Sandimmune formulation. Despite the small numbers of patients and because so many were poor absorbers using Sandimmune, it was possible to demonstrate significant improvements in drug absorption after conversion to Neoral. There was a reduction in the time to maximal blood concentration and an increase in the maximal concentration. The 9-hr cyclosporine AUC showed considerable increases with Neoral, most of which occurred within the first 4 hr after administration of the dose. Within the short-term 12-week follow-up period of this study, no measurable penalty in terms of alteration in renal function was noted. There was a total dose reduction from 4.8 to 4.3 mg/kg/day in the group overall, representing a 10% reduction in drug dispensing to this group of patients and a 17% reduction in mean percentage dose for the individual.
The Neoral formulation was designed in response to the known problems with cyclosporine absorption using the previous Sandimmune formulation(5). Neoral creates a microemulsion on contact with the gastrointestinal contents and was thus designed to enhance both the rate of absorption and bioavailability. Theoretically, absorption of the Neoral formulation should be relatively independent of the bile acid production, which is needed for Sandimmune absorption. Data from liver transplant recipients, both with and without T-tube drainage of bile, have demonstrated improved absorption from Neoral in the presence of bile. However, unlike the experience with Sandimmune, it was possible to maintain acceptable cyclosporine blood concentrations in the early period after transplantation using 7.5 mg/kg/dose of Neoral in the absence of bile acid(6).
Our data agree with studies undertaken in renal transplant recipients, where improved absorption and decreased variability have been demonstrated with the use of Neoral (7-9). The strong and consistent view from Lindholm and Kahan that cyclosporine pharmacokinetic parameters are important variables influencing the outcome of renal transplants (10) has thus been given renewed enthusiasm. Others have supported that view and reconsidered the issue because of the hope that simple and practical measures of blood levels will be more predictive of each individual's absorption profile with Neoral than is the case with Sandimmune (11).
With respect to recipients of simultaneous pancreas and kidney transplants, the conclusions drawn from our study help in the design of protocols for conversion of further patients to Neoral and of protocols for commencing patients de novo after transplantation on Neoral. With respect to conversion protocols, although it would be easier to implement a single dose conversion ratio such as 1:1 or 1:0.9, this denies the observation that there was considerable variability from patient to patient. A more rational approach would be to assess patients before conversion using a reduced AUC sample protocol with blood taken 0, 1, and 2 hr after a test dose. Poor absorbers and underdosed patients will not exceed a concentration of 500 ng/ml in this time period, whereas good absorbers will. In the latter group, a 1:1 conversion ratio can be expected to cause little change in the total quantity of absorbed drug. In the remaining patients, conversion to Neoral can be expected to lead to an increase in the quantity of absorbed drug. There is, however, no alternative but to individualize each patient's dose change on the basis of serum creatinine as an indicator of nephrotoxocity and measured cyclosporine blood levels. Large scale trials and careful multivariate analysis will be needed before one can decide upon which of the pharmacokinetic parameters the Neoral dose should be optimized, because there is at least a theoretically valid basis for selecting any or all of AUC, Cmax, and trough level.
A further protocol that needs to be designed for recipients of simultaneous pancreas and kidney transplants is using Neoral instead of Sandimmune for de novo therapy. The design of this study can only give limited perspectives for de novo as opposed to stable long-term therapy. When compared with current protocols using Sandimmune, it would be reasonable to expect an overall increase in the absorption of cyclosporine for a given dose if Neoral is used instead. It is likely that the pharmacy departments will see at least a 10% reduction in the quantity of drug dispensed, but it is unlikely that this reduction will be uniform across all patients. The implication is that Neoral drug dosing will have to be individualized in the same way that Sandimmune has been. Indeed, it is possible, given the improved pharmacokinetic profile of Neoral, that physicians will design studies to manage patients using parameters other than intermittent trough levels in the hope of improving both the safety and the efficacy of cyclosporine therapy, though the trough level remains the most clinically applicable measure.
There was 1 surprising finding in this study with respect to the changes observed in the GFR. There was a statistically significant negative correlation between a reduction in the GFR and shortening of Tmax. In other words, a larger drop in GFR correlated with a smaller or no drop in Tmax. It is known that continuously measured GFR drops as the blood concentration of cyclosporine rises (12). The protocol of our study ensured that the GFR was actually measured during the first 4 hr after administration of both preparations of cyclosporine, and we thus would have expected, if anything, the reverse, with Neoral providing a greater absorption of cyclosporine in those first 4 hr. It was thus surprising to see this negative correlation, the explanation for which may lie in the fact that it was only observed at 1 time point and only when the data were adjusted for dose. It is now Westmead Hospital practice to withhold for 4 hr the morning cyclosporine dose when measuring the GFR using short-term clearance studies, to avoid potential bias.
In conclusion, this study has confirmed that in previously diabetic recipients of combined renal pancreas allografts, Neoral has superior absorption characteristics compared with Sandimmune, which it will thus largely replace in clinical practice. It is even more important now to determine which of the measurable pharmacokinetic parameters correlate best with immunosuppressive potency and side effects, in the hope that it may be possible with this information to reliably dose patients within the therapeutic window.
Acknowledgments. We thank the staff of the nuclear medicine department for measurement of GFR, the staff of the radioimmunoassay laboratory for measurement of cyclosporine levels, the staff of the pharmacy department for dispensing and monitoring drug utilization, and the physicians who have referred patients to the National Pancreas Transplant Unit included in this study. We thank Dr. Michael Adena, Intstat Australia Pty Ltd. for data entry and validation, and Dr. Karen Byth for her statistical analyses of the data. We also thank Ms. Stephanie Davidson (Sandoz Australia Pty Ltd) for monitoring the trial to Good Clinical Practice standards.
REFERENCES
1. Cheung AHS, Sutherland DER, Gillingham KJ, et al. Simultaneous pancreas-kidney transplant versus kidney transplant alone in diabetic patients. Kidney Int 1992; 41: 924.
2. Munda R, Schroeder TJ, Pedersen SA, et al. Cyclosporine pharmacokinetics in pancreas transplant recipients. Transplant Proc 1988; 20: 487.
3. Kovarik JM, Meuller EA, Johnston A, Hitzenberger G, Kutz K. Bioequivalence of soft gelatin capsules and oral solution of a new cyclosporine formulation. Pharmacotherapy 1993; 13: 613.
4. Fawdry RM, Grunewald SM, Collins LT, Roberts AJ. Comparative assessment of techniques for investigation of glomerular filtration rate with 99mTc-DTPA. Eur J Nucl Med 1985; 11: 7.
5. Vonderscher J, Meinzer A. Rationale for the development of Sandimmune Neoral. Transplant Proc 1994; 26: 2925.
6. Winkler M, Ringe B, Oldhafer K, et al. Influence of bile on cyclosporin absorption for microemulsion formulation in primary liver transplant recipients. Transplant Int 1995; 8: 324.
7. Holt DW, Mueller EA, Kovarik JM, van Bree JB, Kutz K. The pharmacokinetics of Sandimmune Neoral: a new oral formulation of cyclosporine. Transplant Proc 1994; 26: 2935.
8. Kahan BD, Dunn J, Fitts C, et al. The Neoral formulation: improved correlation between cyclosporine trough level and exposure in stable renal transplant recipients. Transplant Proc 1994; 26: 2940.
9. Neumayer HH, Farber L, Haller P, et al. The conversion from Sandimmune to Sandimmune Neoral in patients with stable renal allografts: results after three months. Kidney Int 1995; 47: 990.
10. Lindholm A, Kahan BD. Influence of cyclosporine pharmacokinetics, trough concentrations, and AUC monitoring on outcome after kidney transplantation. Clin Pharmacol Ther 1993; 54: 205.
11. Sketris IS, Lawen JG, Beauregard-Zollinger L, et al. Comparison of the pharmacokinetics of Cyclosporine Sandimmune with Sandimmune Neoral in stable renal transplant patients. Transplant Proc 1984; 26: 2961.
12. Kovarik JM, Kallay Z, Mueller EA, van Bree JB, Arns W, Renner E. Acute effect of cyclosporin on renal function following initial changeover to a microemulsion formulation in stable kidney transplant patients. Transplant Int 1995; 8: 335.
