Increased systemic exposure to statins and consequent risk for complications has been reported in patients concomitantly treated with cyclosporin A (CsA). This has been ascribed to inhibition of drug catabolism by cytochrome P450 3A4 (CYP3A4) or drug transport by P-glycoprotein (PGP) and organic anion transporting polypeptide (OATP1B1). It is not known whether the combination of statins and tacrolimus (Tac) also suffers from this drawback.
Therefore, a pharmacokinetic study of atorvastatin and its metabolites was performed in 13 healthy volunteers after 4 days' treatment, and after short (12 h) concomitant exposure to CsA and Tac. A complementary assessment of overall CYP, and hepatic and intestinal CYP3A4 + PGP activity was performed after each treatment episode and compared to baseline (no drugs). Systemic exposure to atorvastatin acid and its metabolites was significantly increased when administered with CsA. In contrast, intake of Tac did not have any impact on atorvastatin pharmacokinetics. Concomitantly, a profound decrease of hepatic and intestinal PGP and an increase of intestinal CYP3A4 were noted with CsA, whereas no effect was seen after atorvastatin therapy with or without Tac. Based on these findings treatment with Tac appears a safer option for patients needing a combination of statins and calcineurin inhibitors.
Since life- and graft-threatening complications early after transplantation have been largely reduced, cardiovascular disease has become the main cause of death with a functioning graft (1,2). The incidence of dyslipidaemia, a well-known and frequent risk factor for atherosclerotic disease in the general population, is even higher in transplanted patients. Hence not surprisingly, therapy with statins is frequently started in these patients, given the well-established efficacy of statins in the prevention of cardiovascular disease in various patient cohorts, including renal transplant recipients (3–7).
However, an increased frequency of serious adverse events such as severe rhabdomyolysis has been observed in patients treated with cyclosporine A (CsA) and statins (8,9). Subsequent pharmacokinetic studies have shown that the systemic exposure to statins is severalfold increased when coadministered with CsA (10). This has been ascribed to a competitive inhibitory effect of CsA on drug catabolism by cytochrome P450 3A4 (CYP3A4), and efflux by P-glycoprotein (PGP) as well as restricted distribution via other transporters, such as organic anion transport polypeptide (OATP1B1) (11–13). Based on these hypotheses, a similar pattern of interactions has been anticipated with the other frequently used calcineurin inhibitor (CNI) and substrate for CYP3A4, PGP and OATP1B1, tacrolimus (Tac) (14). Consequently, it is generally advocated to reduce the dose of statins, not only in combination with CsA, but also with Tac. However, literature reports on interactions with Tac are much scarcer (15,16).
Recent in vivo experiments on the effects of CNI on CYP3A4 and PGP activity in patients and healthy volunteers demonstrated that only clinical doses of CsA influence these elimination pathways, while Tac has no effect (17). This would suggest that the combination of statins and Tac is safer without the need for dose reduction.
To investigate this, a pharmacokinetic study of atorvastatin and its metabolites was performed in healthy volunteers after atorvastatin therapy with and without short periods of CsA/Tac. Atorvastatin was chosen, since it is a potent statin with a well- documented efficacy in various clinical settings and a rapidly expanding use (18,19). In addition, overall CYP and intestinal/hepatic CYP3A4 and PGP activity was also assessed. Since modulating effects on CYP3A4/PGP by the statins themselves have been reported as well, the activity of these proteins was also measured after therapy with three other frequently used statins: fluvastatin, pravastatin and simvastatin.
Thirteen Caucasian drug-free and nonsmoking healthy males were included in the study, which they all concluded completely. The study had an open-label, randomized, cross-over design as outlined in Figure 1. Baseline biochemistry, consisting of a lipid profile, liver function tests and muscle enzymes, and CYP (3A4)/PGP activity in vivo were determined and compared with: (1) measurements after 4 days' treatment with atorvastatin at a dose of 40 mg daily (2) measurements after a similar treatment with atorvastatin plus clinically relevant doses of CsA/Tac. Pharmacokinetics of atorvastatin without CNI were compared with measurements after coadministration of each CNI. The dose of CsA and Tac amounted to respectively 2.5 mg/kg and 0.0625 mg/kg. To minimize potential toxic side effects, exposure to CsA/Tac was kept short with administration of only two doses (first dose at 12 h before testing, second dose at the time of testing). In addition, a subgroup of eight participants completed a second study, with different statins, as shown in Figure 2.
During the study, all subjects abstained from excessive alcohol consumption (>2 U daily) and avoided known dietary modulators of CYP3A4/PGP, such as grapefruit and related citrus fruits and their juices as well as popular herbal remedies such as St. John's wort. Plasma samples were obtained right before intake of atorvastatin and at 30, 60, 90, 120, 180, 240, 300 and 360 min after drug intake. The Ethics Committee of the University of Leuven approved the study protocol. Written informed consent was obtained from all subjects.
Determination of atorvastatin
Blood samples were drawn on ice, plasma separated by centrifugation at 4°C within minutes and immediately frozen and stored at −80°C until analysis. Sample preparation was performed by solid-phase extraction, followed by separation of the analytes on a HPLC system with a linear gradient and a mobile phase consisting of acetonitrile, water, and formic acid, and detection by electrospray tandem MS (20,21). LOQ was 0.2 ng/mL for atorvastatin and p-hydroxyatorvastatin and 0.5 ng/mL for o-hydroxyatorvastatin. Validation of the analytical method showed linearity from LOQ-600 ng/mL (r2 > 0.99) and the accuracy and precision deviated ≤20.7% in the lower area (20,21). Quality control samples consisting of duplicate spiked plasma samples at 1 and 100 ng/mL were included in each run. The run was accepted when the values of three or more QC samples deviated less than 20% from the nominal concentration. Unfortunately, we did not obtain reference compounds for the lactone forms of atorvastatin, o-hydroxyatorvastatin or p-hydroxyatorvastatin and it was therefore not possible to validate the analytical method with regard to the lactones. Thus, pharmacokinetic data for the lactone compounds are subject to a higher degree of uncertainty than the values of the validated acid forms. Data regarding the lactone forms are presented as arbitrary units (AU).
Area under the plasma concentration versus time curve (AUC) was calculated using the trapezoidal method in the sampling interval 0–6 h (AUC0–6). Trough concentration (C0), peak concentration (Cmax) and time to Cmax (Tmax) were determined from the actually measured values.
Assessment of hepatic and intestinal CYP3A4 and PGP and global CYP activity
Activity of CYP3A4 and PGP was assessed after an overnight fast with the i.v. and per oral 14C-erythromycin breath and urine test (i.v. and po EBT and EUT) as described previously (22). In addition, the 13C aminopyrin breath test (ABT) was used for assessment of global CYP activity (23).
Because of the small number of subjects, nonparametric ANOVA (Friedman) was applied to in vivo measurements of protein activity with correction for multiple testing by calculation of the Rank Least Significant Difference. Pharmacokinetic parameters were analyzed as difference in delta values for the two CNI treatments as compared to atorvastatin alone treatment. Paired Student's t-test on logarithmic transformed data were applied for C0, Cmax and AUC0–6 while Tmax was analyzed by Wilcoxon signed-rank test.
A two-sided p-value <0.05 was considered statistically significant.
Table 1. Pharmacokinetic variables (mean ± SD) of atorvastatin, o-hydroxyatorvastatin and p-hydroxyatorvastatin (acid and lactone forms) in healthy volunteers at baseline (bl), and after treatment with tacrolimus (Tac) and cyclosporin A (CsA)
AUC0–6, area under the plasma concentration versus time curve calculated using the trapezoidal method in the sampling interval (0–6 h); Cmax, maximum plasma concentration; C0, trough plasma concentration; Tmax, time to Cmax;
bl, baseline condition (no concomitant calcineurin inhibitor); Tac, tacrolimus + atorvastatin; CsA, cyclosporin A + atorvastatin.
1Due to lack of reference compounds plasma values are given in arbitrary units.
2Statistically significant from baseline (atorvastatin alone) (p < 0.0001).
3Statistically significant from baseline (atorvastatin alone) (p < 0.01).
4Statistically significant from baseline (atorvastatin alone) (p < 0.05).
67 ± 26
75 ± 42
1026 ± 9062
1.7 ± 1.0
2.4 ± 1.8
10.7 ± 21.74
26.5 ± 13.4
35.2 ± 33.3
362 ± 2652
55 ± 59
53 ± 44
83 ± 284
56 ± 35
59 ± 27
240 ± 1612
2.3 ± 2.0
2.3 ± 1.6
5.4 ± 5.52
12.8 ± 6.9
13.6 ± 6.0
60.2 ± 34.22
173 ± 88
224 ± 95
168 ± 65
11 ± 13
9 ± 7
167 ± 1862
1.0 ± 0.8
1.3 ± 0.9
2.9 ± 2.83
3.1 ± 4.2
2.7 ± 1.8
54 ± 652
226 ± 141
183 ± 149
175 ± 85
12 ± 11
15 ± 12
56.6 ± 50.82
1.0 ± 0.8
1.1 ± 1.0
2.8 ± 2.72
2.9 ± 2.3
3.9 ± 2.9
13.8 ± 11.62
178 ± 86
235 ± 102
210 ± 76
91 ± 62
100 ± 68
361 ± 3432
2.6 ± 1.5
5.5 ± 6.9
7.5 ± 6.72
27.6 ± 19.6
27.1 ± 20.0
96.7 ± 90.62
102 ± 84
152 ± 110
189 ± 1023
72 ± 58
67 ± 44
127 ± 994
2.4 ± 2.0
2.7 ± 2.5
5.2 ± 5.32
16.5 ± 12.6
16.3 ± 9.2
33.9 ± 25.23
212 ± 90
242 ± 83
277 ± 90
Table 2. Pharmacodynamics of atorvastatin at baseline and plus CsA/Tac
No significant difference between the different conditions was noted, except for total cholesterol and LDL-c after all administered drug regimens compared to baseline (*).
Total cholesterol (mg/dL)*
Serum bilirubin (mg/dL)
As can be appreciated from the plasma versus time curves in Figures 3 and 4 and the derived pharmacokinetic variables in Table 1, the exposure to the active, acid, parent compound was increased 15-fold when CsA was administered concomitantly. A significant 4-fold increase was also observed for the lactone form of atorvastatin. Corresponding significant increases were also observed for both the p-hydroxy as well as the o-hydroxy metabolites of atorvastatin acid and lactone. In contrast, coadministration of Tac yielded plasma versus time curves, which were superimposable to those observed after atorvastatin only.
The pharmacodynamic effect of atorvastatin was assessed by its impact on lipid levels and is shown in Table 2. A significant decrease amounting to, respectively, 27% and 39% of total cholesterol and LDL cholesterol at baseline was noticed and this was not influenced by the short concomitant treatment with either of the two CNI. In addition, the level of creatine kinase and liver function parameters were not altered in the presence of CsA or Tac.
Therapy with atorvastatin alone or atorvastatin plus Tac did not influence in vivo assessments of overall CYP, CYP3A4 or PGP. In contrast, when CsA was added to the therapy, a strong decrease of both intestinal and hepatic PGP activity and also a strong increase in intestinal CYP3A4 activity was observed. Overall CYP activity was not altered (Figure 5).
Concerning the effects of the three other statins, only simvastatin produced a moderate, but significant, increase of intestinal CYP3A4 activity. No relevant impact on PGP or overall CYP activity was noted (Figure 6).
More than half of all transplant recipients are currently treated with statins and the number is increasing (24). This is based on accruing evidence that statins effectively lower the frequency of cardiovascular complications in persons at increased cardiovascular risk and with mild-to-moderate impairment of renal function, as is often the case in transplanted patients (3–7).
However, patients with an organ transplant are multimedicated and hence are at an increased risk for potentially deleterious drug interactions. Particularly in patients taking a combination of statins and CsA, the risk for hepatocellular toxicity and rhabdomyolysis due to augmented systemic exposure to the statins and/or their metabolites is increased compared to the general population. Initially this interaction was assumed to result from competitive inhibition of CYP3A4-mediated drug catabolism by CsA. Yet the exposure to non-CYP3A4-dependent statins such as pravastatin and fluvastatin is also increased to similar or even higher levels, suggesting another underlying mechanism. This led to the hypothesis that the observed drug interactions are mediated by an inhibitory effect of CsA on transport proteins such as PGP and the liver specific OATP1B1. Inhibition of the former transporter, especially at the level of the enterocyte and hepatocyte, could increase the oral bioavailability of statins and, at the level of other tissues (e.g. muscle) contribute to the accumulation of potentially toxic parent compounds and metabolites. On the other hand, inhibition of the latter transporter could impair the uptake of statins into the hepatocytes, which are the main distribution site for statins as well as the target organ for their lipid-lowering action. This would explain the relative lack of decrease of lipid levels that could be expected by the high plasma levels of statin in the presence of CsA (25).
Our results corroborate this hypothesis in several ways. First, the pharmacokinetic study demonstrated a major increase of atorvastatin acid exposure and a somewhat smaller increase in its lactone form in the presence of CsA. These results are in agreement with previously published data from renal transplant recipients (10). Second, this was also seen for the CYP3A4-mediated hydroxylated metabolites, which indicates that inhibition of CYP3A4 by CsA is not a dominant mechanism for this interaction. This was also confirmed by the present in vivo measurements of CYP3A4: no effect on hepatic CYP3A4 and even an augmentation of intestinal CYPAA4 activity was noticed. Third, in vivo assessment of both intestinal and hepatic PGP activity was strongly diminished after treatment with CsA. The resulting increase of atorvastatin bioavailability is a plausible mechanism for the increased systemic exposure. Fourth, although we were not able to measure hepatic OATP1B1, the lack of pharmacodynamic effects despite the massively increased systemic exposure strongly suggests that CsA not only affects PGP-mediated drug efflux out of cells but also its specific transport into the hepatocytes, where it normally exerts its cholesterol-lowering effects. Admittedly, an alternative explanation, which unfortunately cannot be rebutted completely, is the rather short duration of the CNI coadministration.
Whereas the observed inhibition of PGP activity in vivo correlates with in vitro data, the enhancement of intestinal CYP3A4 activity is in apparent contradiction with in vitro data (26), which suggest an inhibitory effect. However, these experiments were done on microsomes after very short (<1 h) incubations with CsA. As for omeprazole, which was initially also identified as an inhibitor of CYP3A4, experiments on cell or tissue cultures with longer exposure, reflecting clinical practice more accurately, might reveal the opposite effect (27). In addition, ex vivo studies on human intestine and liver and our own in vivo data have demonstrated that CYP3A4 and PGP regulation and expression are not correlated, despite the similar location of the encoding genes on chromosome 7 and the similar location of their corresponding products in hepatocyte and enterocyte (18,22,28).
Since in vitro experiments with the other frequently used CNI, Tac, have demonstrated similar CYP3A4/PGP/OATP1B1-modulating characteristics as for CsA (13,26), many clinicians tend to apply the same strategy of dose reduction when starting statins in patients taking Tac. Although this preemptive measure might explain why interactions between statins and Tac are less frequently reported, there are also reports that the reduced doses of statins (e.g. pravastatin 10 mg) are less effective in lowering cholesterol levels when combined with Tac (29). Finally, unlike for CsA, therapeutic concentrations for Tac remain well below the concentrations used in the mentioned in vitro studies, which may explain the difference in effect on atorvastatin pharmacokinetics and measured CYP3A4 and PGP activity in vivo.
As a consequence, it is expected that coadministration of Tac and statins in general will not result in a relevant pharmacokinetic interaction. Definite proof that this hypothesis is true for at least atorvastatin was delivered by the present study.
Finally, it was also examined whether other frequently used statins had any sizeable impact on CYP3A4 or PGP in vivo. Only for simvastatin, a moderate induction of intestinal CYP3A4 was noted, which might cause a decreased exposure to exclusively CYP3A4-dependent drugs. However, the exposure to most drugs depends on more than one CYP and also on PGP or other transporters, and potential interactions resulting from the observed effect of simvastatin remain to be investigated.
In conclusion, we demonstrated for the first time that the well-known interaction between CsA and atorvastatin does not occur with coadministration of Tac and atorvastatin. Hence, the combination of Tac and atorvastatin appears safer and does not require dose reduction as is advised for the combination with CsA. Likely this holds also true for the other statins. Definite proof of this, however, requires further investigation.
The ever-continuing support and enthusiasm of the clinical nurses A. Herelixka, H. Wieland, R. Eerdekens, and the excellent technical support of C. Dewit, R. Servaes and L. Swinnen and the other members of the Gastrointestinal Research Center UZ Gasthuisberg is greatly acknowledged.
Siri Johannesen at the School of Pharmacy, University of Oslo is also greatly acknowledged for the support during atorvastatin analyses.