PHARMACOKINETICS AND DRUG DISPOSITION Bioavailabilitv of cvclosporine with concomitant ;ifambin abinistration is rnarkedlv less thankredicted bv h e ~ a t i c enzvrne induction The pharmacokinetics of cyclosporine was studied in six healthy volunteers after administration of the drug orally (10 mglkg) and intravenously (3 mgfkg) with and without concomitant rifampin administration. Both blood and plasma (separated at 37" C) samples were analyzed for cyclosporine concentration. For blood and plasma, respectively, clearances of cyclosporine were calculated to be 0.30 and 0.55 Lhrlkg, values for volume of distribution at steady state were 1.31 and 1.68 Likg, and bioavailabilities were 27% and 33% during the pre-rifampin phase. Post-rifampin phase clearances of cyclosporine were 0.42 and 0.79 Lhrlkg, values for volume of distribution at steady state were 1.36 and 1.35 Llkg, and bioavailabilities were 10% and 9% for blood and plasma, respectively. Rifampin not only induces the hepatic metabolism of cyclosporine but also decreases its bioavailability to a greater extent than would be predicted by the increased metabolism. The decreased bioavailability most probably can be explained by an induction of intestinal cytochrome P450 enzymes, which appears to be markedly greater than the THER 1992;52:453-7.) induction of hepatic metabolism. (CLINPHARMACOL

Marv F. Hebert, PharmD, John P. Roberts, MD, Thomayant Prueksaritanont, PhD, and "kslie Z. ~ & e t ,PhD &n Francisco, Calif: Cyclosporine is an immunosuppressant widely used in the prevention of graft rejection. Data suggest that rifampin induces the metabolism of cyclosporine. Several have noted decreases in cyclosporine trough levels after the initiation of rifampin. Offermann et a].* also noted a dramatic decrease in whole blood oral cyclosporine area under the curve (AUC) in a patient receiving concomitant rifampin therapy. The

observed decreased blood concentrations could be caused by decreased absorption or increased metabolism of cyclosporine; however, until now this issue has not been addressed directly. Drug interaction studies have shown that other agents, such as phenytoin and erythromycin, believed to affect the hepatic metabolism of cyclosporine did in fact also alter its ab~orption.~,~

From the Departments of Pharmacy and Surgery, University of California, San Francisco. Supported in part by grant GM 26691 from the National Institutes of Health, Bethesda, Md. Received for publication February 18, 1992; accepted July 17, 1992. Reprint requests: Mary F. Hebert, PharmD, Division of Clinical Pharmacy, University of California, San Francisco, CA 94 143-0622, 13/1/41079

METHODS Six healthy subjects (five men and one woman [subject 11; age range, 26 to 46 years; weight range, 57 to 100 kg) were studied after each gave informed consent. The study protocol, performed under IND 32,619, was approved by the University of California, San Francisco Committee on Human Research. On the basis of medical histories, physical examinations, and standard biochemical blood and urine tests, subjects

CLIN PHARMACOL THER NOVEMBER 1992

454 Hebe& et al. STUDY DESIGN

DAY

1

2

3

i

4

5

6

7

8

9

10 11 12 13 14 15 16 17

n

A phase I

phase II

Fig. 1. Schematic of study design. A: Days of cyclosporine administration; asterisk (*): 600 mg

rifampin, administered orally at bedtime. were classified as healthy. Subjects were asked to refrain from caffeine and ethanol-containing beverages throughout the study. The study was performed in two phases (Fig. 1). During the initial phase (I), baseline oral and intravenous cyclosporine pharmacokinetics were established. The second phase (11) consisted of oral and intravenous cyclosporine pharmacokinetics with concomitant rifampin administration. On all 4 days of cyclosporine administration, subjects received the same diet for breakfast, lunch, and dinner. The order of cyclosporine route of administration (oral or intravenous) was randomized during phase I, and the same order was maintained in phase 11. During phase I, days for oral (10 mglkg) and intravenous (3 mglkg over 2% hours) cyclosporine administration were separated by 3 nonstudy days. At the completion of phase I, subjects began treatment with rifampin (600 mg orally at bedtime) for 11 days. After 7 days of rifampin administration, phase I1 cyclosporine administration began, with 3 days again separating the oral and intravenous cyclosporine study days. One subject (subject 5) was unable to complete the phase I1 intravenous cyclosporine study day because of nausea, vomiting, and abdominal pain that occurred shortly after the intravenous infusion was started. After all administrations, blood samples (10 ml) were collected at 0, V2, 1, 11/2,2,2V2, 3, 3 % , 4 , 6 , 8, 12, 14, a n d 2 4 hours from an indwelling venous catheter. Both blood and plasma (separated at 37" C) were analyzed for cyclosporine by use of an HPLC m e t h ~ d The . ~ minimum detection limit of the assay was 15 nglml.

Data analysis. The total area under the observed blood or plasma concentration- time curve (AUC) was calculated by use of the linear trapezoidal rule. Clearance (CL = DoseIAUC) for each intravenous infusion was calculated by dividing the dose by the corresponding AUC. Bioavailability (F = (AUC,,a,/AUC,,) X (Dose,,/Dose,,,) and volume of distribution at steady state (Vss = CL X MRT) were calculated by use of noncompartmental techniques7 corrected for duration of infusion [MRT = (AUMCI AUC) - (Infusion duration/2)]. All data are presented as the mean values ? SD, and p < 0.05 with the paired t test is considered to be statistically significant. Although the data from subject 5 are included in Tables I and 11, all mean and SD calculations omit the results from this subject. Clearance and V,, are estimated from intravenous data (Tables I and 11). The predicted impact of changes in clearance on bioavailability resulting from hepatic extraction (FH) can be calculated by use of the equation FH = 1 ER, in which the extraction ratio (ER) is estimated as ER = CLIQ,, assuming that all cyclosporine elimination after intravenous administration is by way of hepatic processes. Hepatic blood flow (Q,) is assumed to be 1500 mllmin in a human who weighs 70 kg. All predictions of FH were done with estimations of hepatic blood flow adjusted for body weight. The expected changes in bioavailability as a result of blood clearance changes alone can then be compared with the actual changes seen in bioavailability estimated from oral and intravenous AUC values and doses. Per-

VOLUME 52 NUMBER 5

Rifampin and cyclospovine bwavailability 45 5

Table I. Pharmacokinetic parameters estimated by use of blood cyclosporine measurements obtained after oral (10 mglkg) and intravenous (3 mglkg) administrations during the pre-rifampin phase I Subject No.

AUCw (ng . hrlml)

A C (ng . hrlml)

Measured F (%)

Predicted F , (%)

1

8,138 10,508 12,718 8,595 10,371 10,499

9,667 4,932 15,100 7,367 5,568 7,866

36 14 36 26 16 22

71 79 81 74 77 78

2 3 4 5 6 Mean

+ SD 10,092 + 1,823 8,986 + 3,813 27 + 9

77

+4

Calculated Fa,, F , (YO) CL (Lihrlkg)

51 18

0.37 0.27 0.24 0.34 0.29 0.28

44

35 21 28 33

+ 13

0.30

V,, (Llkg)

1.67 1.15 1.15 1.34 0.41 1.26

+ 0.05 1.31 + 0.22

AUC. Area under the concentration-time curve; IV, intravenous; F, bioavailability: FH, fraction of the dose absorbed intact into the hepat~cportal vein that i\ not lost to first-pass metabolism; F,,,, fraction of the dose of drug absorbed from the gut lumen; F,, fraction of the dose absorbed that 1s not metabolized by the enzymatic processes in the gut membrane; CL, clearance: V,,, steady-state volume of distribution.

Table 11. Pharmacokinetic parameters estimated by use of blood cyclosporine measurements obtained after oral (10 mglkg) and intravenous (3 mglkg) administrations during the rifampin phase I1 Subject No.

A UC,v (ng . hrlml)

AUC,,,,, (ng . hrlml)

Measured F (%I

Predicted F , (%)

C L (Llhrlkg)

VS, (LIkgl

I

2 3 4 5 6 Mean

k

SD

p Values p Values noted for those values statistically different from phase

I

centage decrease in bioavailability for predicted and actual values were calculated by the difference in the values before and after rifampin administration divided by the prevalue multiplied by 100. The bioavailability of cyclosporine will be affected by at least three factors: the fraction of the dose of drug absorbed from the gut lumen (Fabs), the fraction of the dose absorbed that is not metabolized by the enzymatic processes in the gut membrane (FG), and the fraction of the dose absorbed intact into the hepatic portal vein that is not lost to first-pass hepatic metabolism (F,). The measured bioavailability (F) will be the product of these three terms: F

=

Fa,, . FG ' F,

(1)

Because bioavailability is measured before and after rifampin administration and because FH can be predicted as described above, it is possible to estimate the change in Fa,, . F,, or in FG itself if Fa,, is not affected by rifampin administration.

RESULTS The blood pharmacokinetic parameter estimates of individual subjects are shown for the pre-rifampin phase I (Table I) and the post-rifampin phase I1 (Table 11). Statistically significant decreases were seen in blood cyclosporine pharmacokinetic parameters, including the following: intravenous AUC (10,092 2 1823 ng . hrlml versus 7293 ? 1350 ng . hrlml), oral AUC (8986 ? 3813 ng . hrlml versus 2399 ? 1014 ng . hrlml), and bioavailability (27% ? 9% versus 10% ? 3%) when rifampin therapy was given concomitantly. Cyclosporine blood CL increased significantly (0.30 ? 0.05 Llhrlkg versus 0.42 ? 0.10 Llhrlkg) after rifampin therapy. Similar changes in the estimated pharmacokinetic parameters were seen when plasma concentrations were used. Statistically significant decreases were seen in plasma cyclosporine pharmacokinetic parameters including the following: intravenous AUC (5740 ? 1487 ng . hrlml versus 3962 ? 1221 ng . hrlml), oral AUC (5758 -+ 2 158 ng . hrlml

CLIN PHARMACOL THER NOVEMBER 1992

-

SUBJECT #2 BLOOD (ORAL CYCLOSPORINE)

SUBJECT #2

-

BLOOD

IOOOO 7 (INTRAVENOUS

10000 3

CYCLOSPORINE)

---[II-- PRE-RIFAMPIN PHASE

---i)--

---C

--+ RlFAMPlN PHASE

RlFAMPlN PHASE

PRE-RIFAMPIN PHASE

100

,"

10

, 0

10

20

TlME (HOURS)

1

I

I

I

0

10

20

30

TlME (HOURS)

Fig. 2. Blood cyclosporine (CYA) concentration-time profiles obtained after oral (10 mglkg) and intravenous (3 mglkg) administrations before and with concomitant rifampin administration for subject 2.

versus 1224 + 706 ng . hrlml), and bioavailability (33% k 18% versus 9% k 2%) before and after rifampin therapy, respectively. Cyclosporine plasma clearance was significantly increased (0.55 k 0.20 Llhrlkg versus 0.79 + 0.34 Llhrlkg) when rifampin therapy was added. No significant changes were seen in V,, for blood or plasma as a result of concomitant rifampin administration. Fig. 2 depicts representative (subject 2) blood cyclosporine concentration- time profiles for oral and intravenous administrations. Each graph depicts the cyclosporine curves before and with concomitant rifampin administration. As can be seen in Fig. 2, the relative AUC values before and with rifampin administration show a much greater decrease in AUC with oral compared with intravenous cyclosporine administration. With use of measured blood cyclosporine clearance and estimated hepatic blood flow, one can predict the expected impact of increased clearance on bioavailability. The mean predicted fractional decrease in bioavailability was 12% f- 7 % (i.e., change in mean F, from 77% to 68% in Tables I and 11), whereas the actual fractional decrease was 59% + 21% ( p < 0.006; i.e., change in bioavailability from 27% to 10%) as a result of concomitant rifampin therapy. This calculation assumes that rifampin therapy does not alter hepatic blood flow.

DISCUSSION Cyclosporine is an immunosuppressive agent with well-established usefulness in transplant patients for the prevention of rejection. Unfortunately, one of the complications of long-term immunosuppression is the predisposition to infections such as Mycobacterium tu-

berculosis. Conventional treatment for tuberculosis usually uses a combination of agents, including rifampin. Rifampin is a complex macrocyclic zwitterion for which properties of hepatic enzyme induction are well established. The pharmacokinetics of cyclosporine have been studied previously in both transplant recipients and healthy volunteers. The pharmacokinetic parameters estimated during the pre-rifampin phase of this study are consistent with those previously reported by our laboratory8 and others9 in healthy volunteers. Healthy volunteers were studied to minimize the variables that often make analysis of drug interactions difficult, such as disease states and multiple concomitant medications. The diets of the subjects were controlled on all 4 days of cyclosporine administration to eliminate the variability that dietary fat content places on cyclosporine pharmacokinetics.8 Cyctochrome P450 enzymes are found not only in the liver but also in human intestine.'' It therefore seems possible that agents such as rifampin, which are known to be hepatic mixed-function oxidase inducers," may also induce intestinal P450 enzymes. This induction would be expected to cause a reduction in the bioavailability of agents metabolized by these enzymes to a greater extent than would be predicted by an induction of hepatic enzymes alone. This would lead to a decrease in bioavailability to a much greater degree than would be predicted on the basis of increased clearance, as we found in this study. Recently, Kolars et a1.I2 reported on the metabolism of cyclosporine as it crosses the gastrointestinal tract into the portal vein. That study is supported by our findings.

VOLUME 52 NUMBER 5

Rifanzpin and cyclosporine bioavailability 45 7

By use of equation 1 and the individual values reused during this study. Special thanks to Susan Wong for her excellent technical assistance in processing samples. ported in Tables I and 11, it is possible to estimate the effect of rifampin on the gut enzymes. The mean ? SD ratio calculated as "post" relative to "pre" rifampin References is 2.66 ? administration [(Fa,, . FG),,,/(Fab, . F,),,] Van Buren D. Wideman CA, Ried M, et al. The antag1.32, suggesting that gut metabolism is markedly inonistic effect of rifampin upon cyclosporine bioavailduced, if Fa,, is unaffected by rifampin treatment. ability. Transplant Proc 1984;16:1642-5. Note that a comparison of this mean value, with the Offermann G, Keller F, Molzahn M. Low cyclosporin mean calculated (FH)post/(FH)pre = 1.14 '+ 0.10, apA blood levels and acute graft rejection in renal transpears to suggest that induction of gut P450 has signifplant recipient during rifampin treatment. Am J Nephrol 1985;5:385-7. icantly greater effects on bioavailability than that for Cassidy MJD, Zyl-Smit RV, Pascoe MD, Swanepoel the P450 in the liver. The calculations above could be CR, Jacobson JE. Effect of rifampicin on cyclosporin A misread to suggest that the induction of the gut enblood levels in a renal transplant recipient. Nephron zymes leads to greater variability, but this is not true, 1985;41:207-8. as can be seen in Table I and I1 where coefficients of Rowland M, Gupta SK. Cyclosporin-phenytoin interacvariation of both measured bioavailability values aption: reevaluation using metabolite data. Br J Clin Pharproximate 30%. (The increase in Fa,, . FG with rimacol 1987;24:329-34. fampin treatment is 166% -+ 132%, whereas the inGupta SK, Bakran A, Johnson RWG, Rowland M. Cycrease in FH is 14% ? lo%.) closporin-erythromycin interaction in renal transplant There is another important clinical application to patients. Br J Clin Pharmacol 1989;27:475-81. the findings of this study and those of Kolars et a1.12 Prueksaritanont TP, Koike M, Hoener BA, Benet LZ. Transport and metabolism of cyclosporine in isolated Recently a number of investigators have pointed out rat hepatocytes: the effect of lipids. Biochem Pharmacol that the dosing, and associated costs, of cyclosporine 1992;43:1997-2006. therapy may be reduced by concomitant administraBenet LZ, Galeazzi RL. Noncompartmental determination of the potent P450 inhibitor k e t o c ~ n a z o l e . ' ~ ~ ' ~ tion of the steady-state volume of distribution. J Pharm The practice of adding a second drug, not needed for Sci 1979;68:1071-4. clinical reasons, to modify the pharmacokinetics of Gupta SK, Manfro RC, Tomlanovich SJ, Gambertoglio the primary therapeutic agent is undertaken usually JG, Garovoy MR, Benet LZ. Effect of food on the only with great reluctance because of the wellpharmacokinetics of cyclosporine in healthy subjects documented increase in side effects and drug interacfollowing oral and intravenous administration. J Clin tions that accompany polypharmacy. It had been asPharmacol 1990;30:643-53. sumed that the advantage of concomitant administraPtachcinski RJ, Venkataramanan R. Burckart GJ, et al. Cyclosporine kinetics in healthy volunteers. J Clin tion of ketoconazole would be diminished if a more Pharmacol 1987;27:243-8. bioavailable formulation of cyclosporine could be deWatkins PB, Wrighton SA, Schuetz EG, Molowa DT, veloped. However, the data in this report suggest that Guzelian PS. Identification of glucocorticoid-inducible the maximum bioavailability of cyclosporine, in the cyctochromes P-450 in the intestinal mucosa of rats and absence of an inhibitor, will never approach FH beman. J Clin Invest 1987;80:1029-36. cause of marked metabolism in the gut and that forMiguet JP, Mavier P, Soussy CJ, Dhumeaux D. Inducmulation changes could only yield a limited increase tion of hepatic microsomal enzymes after brief adminisin bioavailability. tration of rifampicin in man. Gastroenterology 1977; In conclusion, the concomitant administration of ri72:924-6. fampin and cyclosporine leads to a statistically signifKolars JC, Awni WM, Merion RM, Watkins PB. Firstpass metabolism of cyclosporine by the gut. Lancet 'icant increase in clearance and decrease in bioavail1991;338:1488-90. ability of cyclosporine. The decreased bioavailability First MR, Schroeder TJ, Alexander JW, et al. Cyis greater than would be predicted by the increased closporine dose reduction by ketoconazole adminiclearance alone. The decreased bioavailability most stration in renal transplant recipients. Transplantation probably can be explained by a more pronounced in1991;s l:365-70. duction of intestinal cytochrome P450 enzymes, in adButman SJ, Wild JC, Nolan PE, et al. Prospective dition to induction of hepatic metabolism. study of the safety and financial benefit of ketoconazole We thank Dr. John G. Gambertoglio for his advice and the generous availability of the clinical research facilities

as adjunctive therapy to cyclosporine after heart transplantation. J Heart Lung Transplant 1991;10:351-8.