Abstract
Background
In a recently described probe drug cocktail for clinically relevant drug transporters containing digoxin, furosemide, metformin and rosuvastatin, mutual interactions were essentially absent except for increases in the systemic exposure of rosuvastatin. To optimize the cocktail, we further examined the dose dependence of the effects of metformin and furosemide on rosuvastatin pharmacokinetics.
Methods
This was a randomized, open label, single center, six-treatment, six-period, six-sequence crossover trial. Eighteen healthy male subjects received 10 mg rosuvastatin as reference treatment and, as test treatments, 10 mg rosuvastatin combined with 10, 50 or 500 mg metformin (T1, T2 and T3) or with 1 or 5 mg furosemide (T4 and T5). Primary pharmacokinetic endpoints were rosuvastatin C max (maximum plasma concentration) and AUC0–tz (area under the plasma concentration–time curve from time zero to the last quantifiable concentration).
Results
The relative bioavailability of rosuvastatin was essentially unchanged when administered with metformin in T1 and T2, but in T3 it increased to 152% for AUC0–tz (90% CI 135–171%) and 154% for C max (90% CI 132–180%). Coadministration with furosemide did not change rosuvastatin relative bioavailability in T4, but in T5 it increased slightly to 116% for AUC0–tz (90% CI 102–132%) and 118% for C max (90% CI 98–142%).
Conclusion
The increased systemic exposure of rosuvastatin when administered as part of the proposed transporter cocktail is most likely attributable to metformin and only to a minor degree to furosemide. Reduction of the doses of metformin and furosemide is expected to eliminate the previously described interaction.
EudraCT no. 2015-003052-46, ClinicalTrials.gov identifier NCT02574845.
Key Points
| The use of a probe drug cocktail is a promising approach to reduce the number of clinical studies required to investigate drug–drug interactions (DDI), and was recently extended for the first time to a cocktail with four probe drugs exclusively selected to investigate potential transporter-based DDIs. |
| In this trial with healthy volunteers, doses of metformin and furosemide were optimized to eliminate their interaction with rosuvastatin in the cocktail. Plasma concentrations after dosing of 10 mg oral rosuvastatin were elevated in the presence of 500 mg metformin or 5 mg furosemide (order: metformin > furosemide), but these interactions could be eliminated by reducing the dose of metformin to 50 or 10 mg and furosemide to 1 mg. |
| Another drug cocktail study, using optimized doses of metformin and furosemide, is currently being evaluated. This cocktail is expected to result in minimal mutual interactions, thus providing a valuable new tool in drug development. |
Introduction
Membrane-bound drug transporters can play a relevant role in drug absorption, distribution and elimination [1, 2]. Modulation of transporter activity by concomitant administration of drugs acting as inhibitors or inducers can lead to drug–drug interactions (DDI) affecting the safety or efficacy of the victim drug [3]. Accordingly, in drug development, thorough investigation of the potential of a new molecular entity to be a victim or a perpetrator of transporter-mediated DDI is recommended by EMA, FDA and PMDA guidelines [4–6]. Regulatory authorities expect an appropriate assessment of the inhibition potential of a new drug towards P-glycoprotein (P-gp), organic anion transporting polypeptides OATP1B1 and OATP1B3, organic anion transporters OAT1 and OAT3, organic cation transporter 2 (OCT2), multidrug and toxin extrusion proteins MATE1 and MATE2-K and breast cancer resistance protein (BCRP) [4–6].
An elegant and efficient approach to reduce the number of clinical studies in a development program is the probe drug cocktail, which can be used to explore the effect of a new molecular entity on the pharmacokinetics of several probe drugs, given together simultaneously in a single trial. Whereas cocktails of probe drugs for cytochrome P450 (CYP)-mediated DDI are well established [7, 8], until very recently no such cocktail has been reported for drug transporters involved in clinically relevant DDI. To fill this gap, we published a Phase I interaction trial [9] that for the first time investigated mutual pharmacokinetic interactions in a probe drug cocktail that is exclusively intended to evaluate the potential for transporter-based DDI. Recently, another group reported a microdose victim probe drug cocktail focused on detection of interactions mediated by inhibition of OATP1B, BCRP, P-gp, and CYP3A, but not aimed at addressing OATs, OCT2 and MATEs, illustrating the high interest in cocktail approaches for evaluation of transporter-mediated DDI [10].
The cocktail investigated for mutual interactions in our previous study contained the following probe drugs and dose levels (principal transporters indicated in parentheses): 0.25 mg digoxin (P-gp), 5 mg furosemide (OAT1 and OAT3), 500 mg metformin (OCT2, MATE1, and MATE2-K) and 10 mg rosuvastatin (OATP1B1, OATP1B3, BCRP) [9]. The design of the in vivo trial took comprehensive preceding in vitro investigations into account [11].
It was found that potentially relevant interactions between the four probe cocktail drugs in vivo were essentially absent, with the exception that rosuvastatin C max (maximum measured concentration in plasma) and AUC0–tz (area under the plasma concentration–time curve from zero to the time of the last quantifiable concentration) were 39 and 43% higher, respectively, when all four probe drugs were given together as a cocktail compared to when rosuvastatin was given alone [9]. This result was unexpected and not predictable based on available in vitro data. Based on an examination of published literature, it was considered that metformin was likely the perpetrator responsible for the increase in rosuvastatin systemic exposure [9, 12]. Also, inhibition of rosuvastatin transport by furosemide, for example via gastrointestinal efflux pumps, could not be ruled out as a potential underlying mechanism. Digoxin, on the other hand, was judged to be an unlikely perpetrator due to its generally low propensity to cause DDI and the low dose of 0.25 mg.
Because a cocktail without mutual interaction between the probe substrates would be preferable [4–6], the present trial was performed to investigate whether metformin or furosemide caused the increase in systemic exposure of rosuvastatin, and whether the interaction could be mitigated by using lower doses of the cocktail components, with the aim of optimizing the probe drug cocktail for investigation of transporter-mediated DDI.
Methods
Subjects
Eighteen healthy male subjects aged 18–55 years and with a body mass index of 18.5–29.9 kg/m2 were eligible to participate in this trial (EudraCT no. 2015-003052-46, ClinicalTrials.gov identifier NCT02574845). Women were not included, to avoid any potential interference of hormonal cycle or hormone-based contraceptives with the trial results. The clinical trial protocol was approved by the Ethics Commission of the State Chamber of Physicians of Baden-Württemberg, Stuttgart, Germany, and the Federal Institute for Drugs and Medicinal Products (BfArM), Bonn, Germany. The trial was conducted in accordance with the principles of the Declaration of Helsinki and the International Council for Harmonization guidelines for Good Clinical Practice. All subjects provided written informed consent.
Trial Objectives, Design and Treatments
The primary trial objective was to determine the relative bioavailability of rosuvastatin when administered orally together with three different oral doses of metformin or two different oral doses of furosemide (test treatments T), compared to rosuvastatin administered alone (reference treatment REF). This was a randomized, open label, single center, six-treatment, six-period, six-sequence crossover trial. Investigational medicinal products were rosuvastatin (Crestor® 10 mg film-coated tablet, AstraZeneca GmbH, Germany), metformin hydrochloride (MetfoLiquid GeriaSan® 500 mg/5 mL oral solution, Infectopharm Arzneimittel und Consilium GmbH, Germany), and furosemide (Lasix® liquidum 10 mg/mL oral solution, Sanofi-Aventis Deutschland GmbH, Germany). The following treatments were to be administered to each of the 18 subjects according to one of the six different randomly assigned treatment sequences, with three subjects per sequence and a washout period of at least 6 days between each individual treatment (based on 5 times the half-life of rosuvastatin, which is approximately 19 h [13]):
Reference (REF): 10 mg rosuvastatin
Test 1 (T1): 10 mg rosuvastatin and 10 mg (0.1 mL) metformin hydrochloride
Test 2 (T2): 10 mg rosuvastatin and 50 mg (0.5 mL) metformin hydrochloride
Test 3 (T3): 10 mg rosuvastatin and 500 mg (5 mL) metformin hydrochloride
Test 4 (T4): 10 mg rosuvastatin and 1 mg (0.1 mL) furosemide
Test 5 (T5): 10 mg rosuvastatin and 5 mg (0.5 mL) furosemide
After an overnight fast of at least 10 h, the trial medication was administered to the subjects in a standing position as a single oral dose together with 240 mL of water. During the first 4 h after medication administration, no food was allowed and subjects were not allowed to lie down.
Pharmacokinetics
Blood samples (4.9 mL) for the measurement of plasma concentrations of rosuvastatin, metformin and furosemide were taken using K3-EDTA as anticoagulant from a forearm vein of each subject before dosing and at 30 min, 1 h, 1.5 h, 2 h, 3 h, 4 h, 5 h, 6 h, 8 h, 10 h, 12 h, 24 h, 34 h and 48 h after dosing. Pharmacokinetic parameters were calculated using standard noncompartmental methods with the software Phoenix WinNonlin® Professional, version 6.3 (Certara, Princeton, USA).
The primary endpoints were AUC0–tz and C max for rosuvastatin, and the secondary endpoint was AUC0–∞ for rosuvastatin (area under the plasma concentration–time curve extrapolated to infinite time). Further endpoints were: (1) for rosuvastatin, %AUCtz–∞ (percentage of AUC0–∞ obtained by extrapolation), t max (time to attainment of maximum plasma concentration), t 1/2 (terminal elimination half-life), CL/F (apparent plasma clearance), V z/F (apparent volume of distribution during the terminal elimination phase), and MRTpo (mean residence time); (2) for metformin and furosemide, AUC0–tz and C max; in addition AUC0–∞ for metformin only. The t 1/2 calculation of rosuvastatin was based on at least the last 3 plasma concentration time points (24, 34 and 48 h).
Bioanalytical Methods
Plasma concentrations of rosuvastatin, metformin, and furosemide were determined by fully validated [14, 15] HPLC–MS/MS (high-performance liquid chromatography–tandem mass spectrometry) methods using the isotope-labeled internal standards [13C6]furosemide, [13C1,2H4]rosuvastatin and [2H6]metformin. The assay for simultaneous determination of rosuvastatin (0.2–20 nmol/L) and furosemide (0.6–1500 nmol/L) was performed at Boehringer Ingelheim, Drug Metabolism & Pharmacokinetics, Biberach, Germany. It comprised sample clean-up by supported liquid extraction on Biotage Isolute 200 µL SLE+ 96 well plates with isopropyl acetate as extraction medium followed by chromatography on an analytical Waters xBridge BEH C8 2.5 µm analytical column. Monitored MS/MS transitions on a Sciex 6500QTrap mass spectrometer were from m/z 482.2 → 258.1 and 487.2 → 263.0 for rosuvastatin and its internal standard (positive ion mode) as well as from m/z 328.8 → 204.8 and 335.0 → 211.1 for furosemide and its internal standard (negative ion mode).
The assays for determination of metformin in the low (0.6–300 nmol/L) and high (38.7–19,400 nmol/L) concentration ranges were performed at SGS Cephac Europe, Saint Benoît Cedex, France. They comprised sample clean-up by solid phase extraction on Waters Seppack C18 extraction cartridges with MeOH/MeCN (50/50, v/v) containing 1% of formic acid as elution medium followed by chromatography on analytical Phenomenex Kinetex C18 2.6 µm (low) or Synergi 4 µm polar RP 80A (high) analytical columns. Monitored MS/MS transitions on a Sciex API 4000 (3000) mass spectrometer were from m/z 130.0 (130.2) → 71.1 and 136.1 (136.2) → 77.1 for metformin and its internal standard (positive ion mode).
Safety Assessment
Safety and tolerability were assessed based on adverse events (AE), safety laboratory, vital signs, and 12-lead ECGs. The treated set (consisting of all subjects treated with at least one study drug) was used for safety analyses. AEs were analyzed according to the concept of treatment-emergent AEs and the number of subjects with AEs. AEs occurring within 4 days after study drug intake were defined as treatment-emergent and assigned to the corresponding treatment. The Medical Dictionary for Regulatory Activities (MedDRA) version 18.1 was used for coding of AEs.
Statistical Methods
The relative bioavailability of rosuvastatin administered together with the different doses of metformin and furosemide (“test” treatments T1, T2, T3, T4, T5) compared to rosuvastatin given alone (“reference” treatment REF) was investigated based on the geometric mean ratios (GMR) of test/reference and their two-sided 90% confidence intervals (CI) for the primary (AUC0–tz, C max) and secondary (AUC0–∞) endpoints. The statistical model used was analysis of variance (ANOVA) on the logarithmic scale and included the effects sequence, subjects within sequence, period and treatment. The effect ‘subjects within sequence’ was defined as random, whereas the other effects were fixed. The analysis set included all treated subjects who had at least one evaluable primary or secondary endpoint. The statistical methods of the primary analysis were further used to assess the proportionality of AUC0–tz, C max and AUC0–∞ among the different doses of metformin and furosemide used in the trial. No corrections were performed for multiple comparisons, due to the exploratory nature of the trial. All statistical calculations were performed with SAS® (version 9.4.by SAS Institute Inc., Cary, NC, USA).
The sample size was based on the expected precision (defined as ratio of upper 90% CI limit over point estimate) of the GMR estimate and the within-subject gCVs (geometric coefficients of variation) of the primary endpoints observed in [9]. The calculation was performed as described by Kupper and Hafner [16] using R Version 3.0.3. Assuming an intra-individual gCV of 20% and 18 evaluable subjects, the precision was expected to be 1.16 and deemed sufficient for this trial. Based on these assumptions and a GMR of 100%, the 90% CI was expected to range from 86 to 116%. The dropout rate was expected to be considerably less than 6, the number of investigated treatment sequences. To avoid an unbalanced design in case of no dropouts, it was not planned to recruit additional subjects to account for possible dropouts.
Results
Subjects
Eighteen healthy white male subjects were randomized in the trial and treated. Their mean (SD; standard deviation) age was 31.9 (10.0) years and the mean (SD) body mass index was 26.03 (2.41) kg/m2. Of the 18 treated subjects, 15 (83.3%) completed the planned observation time according to the protocol. One subject withdrew consent after the first treatment for personal reasons. Two subjects were withdrawn due to AEs that were not treatment-emergent or drug-related. Additionally, two subjects missed the second of the six treatments (not due to AE). For one subject, metformin administration in treatment T2 (rosuvastatin + 50 mg metformin) was incomplete because several drops of the oral solution were lost during dosing. The pharmacokinetic data of the subject for that treatment were not used for assessment of the primary endpoints.
Pharmacokinetics
Bioanalytical Methods
Assay performance during measurement of the trial samples was assessed by back-calculation of calibration standards, tabulation of the standard curve fit function parameters and measurement of quality control samples. Accuracy, measured as the deviation from target concentrations, ranged from −3.2 to −5.6% for rosuvastatin, −2.5 to −4.2% for furosemide, and −2.0 to 2.2% (low range), −1.7 to 1.2% (high range) for metformin. The assay precision (coefficient of variation) was 5.9–9.0% for rosuvastatin, 5.7–7.7% for furosemide, and 2.1–5.3% (low range), 1.5–2.9% (high range) for metformin. The results documented adequate accuracy, precision and specificity of the assays.
Effect of Metformin on Rosuvastatin Pharmacokinetics
Geometric mean plasma concentration–time profiles of rosuvastatin 10 mg when given alone (reference treatment) or together with increasing doses of metformin (test treatments T1–T3) are shown in Fig. 1. The pharmacokinetic parameters of rosuvastatin are described in Tables 1 and 2. Maximum plasma concentrations of rosuvastatin were observed after a median t max of 5 h in all treatment periods (Table 2). The profiles of rosuvastatin alone and together with metformin 10 mg (T1) and 50 mg (T2) were almost superimposable, whereas rosuvastatin plasma concentrations were clearly higher when given in combination with metformin 500 mg (T3) (Fig. 1). The relative bioavailabilities of rosuvastatin with metformin 10 mg (T1) and 50 mg (T2) compared to rosuvastatin alone were close to 100% for AUC0–tz, C max and AUC0–∞, and all 90% CI included 100% and were in the range of 87–124%. However, the relative bioavailability of rosuvastatin with respect to these parameters increased to 146–154% when given with metformin 500 mg (T3), and all 90% CIs were in the range of 132–180% (Table 1). The precision of the AUC0–∞values was considered high, with %AUCtz–∞ values of around 10% or below (Table 2)
Fig. 1.
Geometric mean plasma concentration–time profiles of rosuvastatin (10 mg) after oral dosing alone (reference), and in combination with different doses of metformin a linear and b semi-log. The doses of metformin were 10 mg (T1), 50 mg (T2) and 500 mg (T3)
Table 1.
Geometric means and relative bioavailabilities for the primary and secondary pharmacokinetic endpoints of oral rosuvastatin 10 mg given alone (reference treatment REF) or together with different doses of metformin (test treatments T1–T3)
| Endpoint | Test (T) | Reference (REF) | Ratio T/REF (%) | 90% CI (%) | gCV (%)a | ||
|---|---|---|---|---|---|---|---|
| Rosuvastatin + 10 mg metformin (T1) | Rosuvastatin alone | ||||||
| N | Adj. gMean | N | Adj. gMean | ||||
| AUC0–tz (nmol·h/L) | 15 | 85.96 | 16 | 84.32 | 101.95 | (89.49; 116.15) | 19.3 |
| C max (nmol/L) | 15 | 9.54 | 16 | 9.31 | 102.47 | (87.19; 120.42) | 24.1 |
| AUC0–∞ (nmol·h/L) | 14 | 101.72 | 15 | 100.50 | 102.22 | (89.23; 114.82) | 17.8 |
| Endpoint | Test (T) | Reference (REF) | Ratio T/REF (%) | 90% CI (%) | gCV (%)a | ||
|---|---|---|---|---|---|---|---|
| Rosuvastatin + 50 mg metformin (T2) | Rosuvastatin alone | ||||||
| N | Adj. gMean | N | Adj. gMean | ||||
| AUC0–tz (nmol·h/L) | 16 | 88.54 | 16 | 83.45 | 106.09 | (96.12; 117.11) | 14.6 |
| C max (nmol/L) | 16 | 9.90 | 16 | 9.25 | 106.98 | (92.54; 123.69) | 21.7 |
| AUC0–∞ (nmol·h/L) | 16 | 103.96 | 15 | 96.93 | 107.26 | (97.65; 117.81) | 12.6 |
| Endpoint | Test (T) | Reference (REF) | Ratio T/REF (%) | 90% CI (%) | gCV (%)a | ||
|---|---|---|---|---|---|---|---|
| Rosuvastatin + 500 mg metformin (T3) | Rosuvastatin alone | ||||||
| N | Adj. gMean | N | Adj. gMean | ||||
| AUC0-tz (nmol·h/L) | 15 | 128.94 | 16 | 84.73 | 152.18 | (135.12; 171.41) | 18.1 |
| C max (nmol/L) | 15 | 14.48 | 16 | 9.40 | 154.07 | (131.70; 180.24) | 24.0 |
| AUC0–∞ (nmol·h/L) | 14 | 147.48 | 15 | 100.71 | 146.45 | (135.21; 158.62) | 11.4 |
AUC 0–tz area under the plasma concentration–time curve from time zero to last quantifiable concentration, C max peak plasma concentration, AUC 0–∞ area under the plasma concentration–time curve extrapolated to infinity, CI confidence interval, gCV geometric coefficient of variation, Adj. gMean adjusted geometric mean
aWithin-subject
Table 2.
Additional pharmacokinetic parameters of oral rosuvastatin 10 mg given alone (reference treatment) or together with different doses of metformin (test treatments T1–T3)
| Rosuvastatin alone | Rosuvastatin + 10 mg metformin (T1) | |||||
|---|---|---|---|---|---|---|
| N | gMean | gCV (%) | N | gMean | gCV (%) | |
| t max (h)a | 16 | 5.00 | 3.00–5.03 | 15 | 5.00 | 3.00–5.03 |
| t 1/2 (h) | 15 | 24.6 | 54.2 | 14 | 23.4 | 57.3 |
| V z/F (L) | 15 | 7390 | 87.1 | 14 | 6650 | 82.5 |
| CL/F (mL/min) | 15 | 3470 | 39.0 | 14 | 3290 | 48.0 |
| %AUCtz–∞ (%) | 15 | 10.6 | 77.2 | 14 | 9.85 | 74.0 |
| MTRpo (h) | 15 | 20.3 | 56.6 | 14 | 19.1 | 61.5 |
| Rosuvastatin + 50 mg metformin (T2) | Rosuvastatin + 500 mg metformin (T3) | |||||
|---|---|---|---|---|---|---|
| N | gMean | gCV (%) | N | gMean | gCV (%) | |
| t max (h)a | 16 | 5.00 | 1.50–6.00 | 15 | 5.00 | 4.00–5.02 |
| t 1/2 (h) | 16 | 23.0 | 38.4 | 14 | 17.5 | 32.1 |
| V z/F (L) | 16 | 6790 | 84.8 | 14 | 3520 | 80.5 |
| CL/F (mL/min) | 16 | 3410 | 55.0 | 14 | 2330 | 49.8 |
| %AUCtz–∞ [%] | 16 | 10.1 | 76.6 | 14 | 6.09 | 60.0 |
| MRTpo (h) | 16 | 18.6 | 40.7 | 14 | 15.1 | 21.2 |
gMean geometric mean, CL/F apparent plasma clearance, gCV geometric coefficient of variation, t 1/2 terminal elimination half-life, t max time to maximum plasma concentration, V z /F apparent volume of distribution during the terminal elimination phase; %AUC tz–∞ percentage of AUC 0–∞ obtained by extrapolation, MRT po mean residence time of the analyte in the body after oral administration
aMedian and range
Effect of Furosemide on Rosuvastatin Pharmacokinetics
Geometric mean plasma concentration–time profiles of rosuvastatin 10 mg when given alone (reference treatment) or together with increasing doses of furosemide (test treatments T4 and T5) are shown in Fig. 2. The pharmacokinetic parameters of rosuvastatin are described in Tables 3 and 4. Maximum plasma concentrations of rosuvastatin were again observed after a median t max of 5 h in all treatment periods (Table 4). There was no change in plasma concentrations of rosuvastatin with furosemide 1 mg (T4) and only a slight increase with furosemide 5 mg (T5) (Fig. 2). The relative bioavailability of rosuvastatin with respect to AUC0–tz, C max and AUC0–∞ was 106–107% with furosemide 1 mg (T4), and increased only slightly to 116–118% with furosemide 5 mg (T5). With furosemide 1 mg, the 90% CIs of the GMRs included 100% and ranged from 92 to 124%, but with furosemide 5 mg the 90% CI included 100% only for C max. For the other endpoints, the lower limit of the 90% CI was already greater than 100% (Table 3). The precision of the AUC0–∞values was considered high, with %AUCtz–∞ values of around 10% (Table 4).
Fig. 2.
Geometric mean plasma concentration–time profiles of rosuvastatin (10 mg) after oral dosing alone (reference), and in combination with different doses of furosemide a linear and b semi-log. The doses of furosemide were 1 mg (T4) and 5 mg (T5)
Table 3.
Geometric means and relative bioavailabilities for the primary and secondary pharmacokinetic endpoints of rosuvastatin 10 mg given alone (reference treatment REF) or together with different doses of furosemide (test treatments T4 and T5)
| Endpoint | Test (T) | Reference (REF) | Ratio T/REF (%) | 90% CI (%) | gCV (%)a | ||
|---|---|---|---|---|---|---|---|
| Rosuvastatin + 1 mg furosemide (T4) | Rosuvastatin alone | ||||||
| N | Adj. gMean | N | Adj. gMean | ||||
| AUC0–tz (nmol·h/L) | 15 | 90.28 | 16 | 84.40 | 106.97 | (94.34; 121.30) | 18.9 |
| C max (nmol/L) | 15 | 9.94 | 16 | 9.31 | 106.81 | (91.78; 124.30) | 22.9 |
| AUC0–∞ (nmol·h/L) | 15 | 106.40 | 15 | 100.73 | 105.63 | (92.20; 121.00) | 20.4 |
| Endpoint | Test (T) | Reference (REF) | Ratio T/REF | 90% CI | gCV (%)a | ||
|---|---|---|---|---|---|---|---|
| Rosuvastatin + 5 mg furosemide (T5) | Rosuvastatin alone | ||||||
| N | Adj. gMean | N | Adj. gMean | ||||
| AUC0–tz (nmol·h/L) | 16 | 94.29 | 16 | 81.34 | 115.92 | (101.93; 131.82) | 19.7 |
| C max (nmol/L) | 16 | 10.65 | 16 | 9.02 | 117.98 | (98.27; 141.65) | 28.4 |
| AUC0–∞ (nmol·h/L) | 16 | 113.43 | 15 | 96.04 | 118.10 | (106.75; 130.67) | 15.4 |
AUC 0–tz area under the plasma concentration–time curve from time zero to last quantifiable concentration, C max peak plasma concentration, AUC 0–∞ area under the plasma concentration–time curve extrapolated to infinity, CI confidence interval, gCV geometric coefficient of variation, Adj. gMean adjusted geometric mean
aWithin-subject
Table 4.
Additional pharmacokinetic parameters of oral rosuvastatin 10 mg given alone (reference treatment) or together with different doses of furosemide (test treatments T4 and T5)
| Rosuvastatin alone | Rosuvastatin + 1 mg furosemide (T4) | Rosuvastatin + 5 mg furosemide (T5) | |||||||
|---|---|---|---|---|---|---|---|---|---|
| N | gMean | gCV (%) | N | gMean | gCV (%) | N | gMean | gCV (%) | |
| t max (h)a | 16 | 5.00 | 3.00–5.03 | 15 | 5.00 | 3.00–5.05 | 16 | 5.00 | 2.03–5.02 |
| t 1/2 (h) | 15 | 24.6 | 54.2 | 15 | 23.0 | 61.8 | 16 | 25.1 | 61.8 |
| V z/F (L) | 15 | 7390 | 87.1 | 15 | 6390 | 90.6 | 16 | 6530 | 93.8 |
| CL/F (mL/min) | 15 | 3470 | 39.0 | 15 | 3210 | 47.9 | 16 | 3000 | 42.0 |
| %AUCtz–∞ (%) | 15 | 10.6 | 77.2 | 15 | 9.81 | 92.5 | 16 | 10.8 | 83.5 |
| MRTpo (h) | 15 | 20.3 | 56.6 | 15 | 19.6 | 69.3 | 16 | 20.6 | 69.7 |
gMean geometric mean, CL/F apparent plasma clearance, gCV geometric coefficient of variation, t 1/2 terminal elimination half-life, t max time to maximum plasma concentration, V z /F apparent volume of distribution during the terminal elimination phase, %AUC tz–∞ percentage of AUC 0–∞ obtained by extrapolation, MRT po mean residence time of the analyte in the body after oral administration
aMedian and range
Metformin Pharmacokinetics
The pharmacokinetic parameters AUC0–tz, C max and AUC0–∞ of metformin at the doses of 10, 50 and 500 mg administered together with 10 mg rosuvastatin (test treatments T1–T3) and their pairwise relative bioavailabilities are shown in Table 5. On increasing the metformin dose fivefold (from 10 to 50 mg), tenfold (from 50 to 500 mg) and 50-fold (from 10 to 500 mg), the pharmacokinetic parameters showed less than proportional increases; these were 4.1-, 7.9-, and 33-fold respectively for AUC0–tz, and slightly lower factors for C max.
Table 5.
Dose proportionality assessment for pharmacokinetic parameters of metformin at three doses (test treatments T1–T3) in the presence of 10 mg rosuvastatin
| Endpoint | Test (T) | Reference (REF) | Ratio T/REF | 90% CI | Dose ratio | (T/REF) as a fraction of dose ratio | ||
|---|---|---|---|---|---|---|---|---|
| T2 (50 mg metformin) | T1 (10 mg metformin) | |||||||
| N | Adj. gMean | N | Adj. gMean | |||||
| AUC0–tz (nmol·h/L) | 16 | 6382.07 | 15 | 1554.00 | 4.11 | (3.87; 4.36) | 5 | 0.822 |
| C max (nmol/L) | 16 | 990.39 | 15 | 273.27 | 3.62 | (3.16; 4.15) | 5 | 0.724 |
| AUC0–∞ (nmol·h/L) | 16 | 6418.12 | 12 | 1559.11 | 4.12 | (3.83; 4.43) | 5 | 0.824 |
| Endpoint | Test (T) | Reference (REF) | Ratio T/REF | 90% CI | Dose ratio | (T/REF) as a fraction of dose ratio | ||
|---|---|---|---|---|---|---|---|---|
| T3 (500 mg metformin) | T2 (50 mg metformin) | |||||||
| N | Adj. gMean | N | Adj. gMean | |||||
| AUC0–tz (nmol·h/L) | 15 | 50027.10 | 16 | 6333.14 | 7.90 | (7.55; 8.27) | 10 | 0.790 |
| C max (nmol/L) | 15 | 7129.32 | 16 | 999.06 | 7.14 | (6.54; 7.78) | 10 | 0.714 |
| AUC0–∞ (nmol·h/L) | 15 | 50491.90 | 16 | 6374.86 | 7.92 | (7.58; 8.28) | 10 | 0.792 |
| Endpoint | Test (T) | Reference (REF) | Ratio T/REF | 90% CI | Dose ratio | (T/REF) as a fraction of dose ratio | ||
|---|---|---|---|---|---|---|---|---|
| T3 (500 mg metformin) | T1 (10 mg metformin) | |||||||
| N | Adj. gMean | N | Adj. gMean | |||||
| AUC0–tz (nmol·h/L) | 15 | 50547.30 | 15 | 1553.37 | 32.54 | (31.12; 34.03) | 50 | 0.651 |
| C max (nmol/L) | 15 | 7193.55 | 15 | 266.59 | 26.98 | (23.95; 30.40) | 50 | 0.540 |
| AUC0–∞ (nmol·h/L) | 15 | 51017.10 | 12 | 1537.88 | 33.17 | (31.24; 35.22) | 50 | 0.663 |
AUC 0–tz area under the plasma concentration–time curve from time zero to last quantifiable concentration, C max peak plasma concentration, AUC 0–∞ area under the plasma concentration–time curve extrapolated to infinity, CI confidence interval, Adj. gMean adjusted geometric mean
Furosemide Pharmacokinetics
AUC0–tz and C max of furosemide at the doses of 1 and 5 mg administered together with 10 mg rosuvastatin (test treatments T4 and T5) and the relative bioavailability of the 5 mg compared to the 1 mg dose are shown in Table 6. The GMRs of furosemide between the 5 and 1 mg doses were 5.3 (AUC0–tz) and 5.2 (C max) indicating dose proportionality.
Table 6.
Dose proportionality assessment for pharmacokinetic parameters of furosemide at two doses (test treatments T4 and T5) in the presence of rosuvastatin
| Endpoint | Test (T) | Reference (REF) | Ratio T/REF | 90% CI | Dose ratio | (T/REF) as a fraction of dose ratio | ||
|---|---|---|---|---|---|---|---|---|
| T5 (5 mg furosemide) | T4 (1 mg furosemide) | |||||||
| N | Adj. gMean | N | Adj. gMean | |||||
| AUC0–tz (nmol·h/L) | 16 | 980.05 | 15 | 183.59 | 5.34 | (4.91; 5.80) | 5 | 1.068 |
| C max (nmol/L) | 16 | 493.03 | 15 | 94.72 | 5.21 | (4.53; 5.98) | 5 | 1.042 |
AUC 0–tz area under the plasma concentration–time curve from time zero to last quantifiable concentration, C max peak plasma concentration, CI confidence interval, Adj. gMean adjusted geometric mean
Safety
Treatment-emergent AEs were reported by 9/18 subjects (50.0%). One subject reported mild or moderate headache that was assessed as drug-related by the investigator after REF, T1, and T2. No other AEs were assessed as drug-related. Two subjects discontinued the study due to AEs not defined as treatment-emergent and assessed as not drug-related. One subject was diagnosed with asymptomatic Wolff–Parkinson–White syndrome, a congenital ECG disorder, before treatment in period 2. In ECGs before this time point, the abnormality had been hidden. Another subject reported severe back pain that caused him to drop out after period 4. No deaths, other treatment-emergent SAEs, or ‘other significant AEs’ occurred during the trial. There were no treatment-emergent clinically relevant findings in clinical laboratory, vital signs, or ECG. Rosuvastatin, metformin and furosemide were safe and well tolerated in this trial.
Discussion
The pharmacokinetic data obtained in this trial confirm and extend the results presented in our previous report, which was the first clinical investigation of mutual pharmacokinetic interactions within a dedicated drug transporter cocktail containing 0.25 mg digoxin, 5 mg furosemide, 500 mg metformin and 10 mg rosuvastatin [9]. In that initial trial, two additional dosing regimens were also studied that contained higher metformin and furosemide doses (1000 and 20 mg, respectively) to simulate conditions in which these drugs could act as perpetrators of drug-transporter interactions [9]. It was found that the most pronounced pharmacokinetic change identified in the trial—an approximate 40% increase in rosuvastatin C max and AUC compared to dosing alone—did not increase further at the higher doses of metformin and furosemide. In the present trial, the metformin and furosemide dose ranges were adjusted to substantially lower levels—from 500 to 50 mg and 10 mg for metformin and from 5 to 1 mg for furosemide—in order to establish “no-effect” doses for these compounds with respect to their influence on the AUC and C max of rosuvastatin.
The increase in AUC0–tz, C max and AUC0–∞ of 10 mg rosuvastatin by 46–54% after coadministration with 500 mg metformin (T3), Table 1, was comparable to that observed after coadministration with 500 mg metformin, 5 mg furosemide and 0.25 mg digoxin in the initial trial, where AUC0–tz, C max and AUC0–∞ values increased by 43, 39 and 28%, respectively [9].
Reduction of the dose of metformin to 50 mg (T2) and 10 mg (T1) in the current trial was sufficient to show similarity of rosuvastatin AUC0–tz, C max and AUC0–∞ values (Table 1). Additional pharmacokinetic parameters were also not relevantly changed (Table 2). It is therefore feasible to reduce the metformin dose in the original cocktail from 500 to 50 mg or 10 mg to eliminate this interaction. There was a slight increase in AUC0–tz, C max and AUC0–∞ of 10 mg rosuvastatin by 16–18% in this trial with coadministration of 5 mg furosemide (T5) as shown in Table 3. Here also, further reduction of the furosemide dose to 1 mg (T4) resulted in similarity of the rosuvastatin pharmacokinetic parameters, with relative bioavailabilities of AUC and C max (GMRs) around 107% and all 90% CIs including 100% (Table 3), and additional pharmacokinetic parameters were not relevantly changed (Table 4). The furosemide dose in the original cocktail could therefore be reduced from 5 to 1 mg to effectively eliminate this slight interaction. Taken together, the present results indicate that the increase in rosuvastatin bioavailability in the cocktail in the initial trial [9] was most likely principally caused by metformin, but with a small contribution by furosemide.
A relatively minor interaction of metformin with rosuvastatin was also reported by Lee et al. [12] in a steady-state trial in Asian subjects with 750 mg metformin and 10 mg rosuvastatin q.d. In that study, the AUCτ,ss of rosuvastatin was hardly affected by coadministration of metformin (11% increase, CIs within the acceptance limits), and C max,ss increased slightly by 23% (CIs outside the acceptance limits). These changes—in common with those we observed in this trial and the previous report [9]—are very small in comparison to those caused by rifampin, which is a potent inhibitor of OATP1B1, OATP1B3 and an inhibitor of BCRP [10, 17]. To our knowledge, no studies have been published to date on any clinical interaction of furosemide with rosuvastatin.
In our previous trial [9], the increase of the systemic rosuvastatin exposure in the cocktail by approximately 40% was accompanied by an increase in the fraction of rosuvastatin excreted in urine without a relevant change in its renal clearance. This indicates that metformin and furosemide interact with rosuvastatin on an extrarenal mechanism. The molecular mechanisms underlying the increased AUC and C max of rosuvastatin caused by the higher doses of metformin and furosemide in this trial require consideration of the known transporters for which rosuvastatin is a substrate. Based on in vitro data, Ebner et al. assessed the risk for interaction within the cocktail at the level of OATP1B- and BCRP-mediated rosuvastatin transport as remote [11]. In addition, rosuvastatin transport was reported for multidrug resistance protein 2 (MRP2) [18, 19], OATP2B1 (transport data are controversial) [18, 20, 21] and sodium taurocholate cotransporting polypeptide (NTCP) [22, 23]. However, the risk for mutual interaction within the cocktail at these transporters is also regarded as remote (Boehringer Ingelheim, data on file). Based on the available data, the precise molecular mechanisms underlying the interactions of rosuvastatin with metformin and furosemide are currently unknown, although mechanisms affecting intestinal absorption or intestinal metabolism may be involved that would require further in vitro and in vivo work for clarification.
This trial also allowed the assessment of the dose proportionality of the pharmacokinetics of metformin and furosemide in the presence of 10 mg rosuvastatin. Metformin exhibited less-than-proportional increases of AUC and C max values with dose (Table 5), similar to the observations in the initial trial using doses of 500 and 1000 mg [9]. Kusuhara et al. reported that AUC0–12h of metformin showed dose proportionality when comparing a microdose of 100 μg and a therapeutic dose of 250 mg in healthy subjects [24], but the dose range used was different and blood sampling was not as extended as in the present trial. Metformin dose non-proportionality could be explained by a decreasing extent of absorption with increasing dose [25]. Non-dose proportional absorption would not be expected to affect the sensitivity of the cocktail for detection of clinically relevant inhibition of OCT2 or MATEs localized in the kidney. The dose-proportionality of furosemide pharmacokinetics observed between 1 and 5 mg (Table 6) is in agreement with and extends literature data describing dose proportionality in the range of 20–80 mg [26].
Another observation in the original cocktail was a decrease of furosemide C max by 19% as compared to furosemide 5 mg alone without a change in furosemide renal clearance, suggesting an extrarenal mechanism [9]. This minor and presumably extrarenal interaction is not expected to affect the sensitivity of the cocktail for inhibition of renal transporters OAT1 and OAT3; however, a fully interaction-free cocktail would be preferable. Our finding is consistent with a previous report of a 31% decrease of furosemide C max when given together with metformin [27]. If indeed metformin is the perpetrator of this relatively minor interaction, a 10- or 50-fold decrease of metformin dose from 500 mg in the original cocktail to 50 or 10 mg could, as a positive secondary effect, free the cocktail from significant changes in furosemide plasma concentrations.
On the basis of the new results gained in this trial, it is proposed that the four-component drug transporter cocktail [9] be re-evaluated and further clinically tested using the doses 0.25 mg digoxin, 1 mg furosemide (previously 5 mg), 10 or 50 mg metformin (previously 500 mg) and 10 mg rosuvastatin. Such a trial is ongoing and will determine the pharmacokinetics in plasma and urine of each of the components alone and in presence of the other three and is expected to show minimal mutual interactions.
Conclusion
The increased exposure of rosuvastatin that was observed when administered as part of the proposed transporter cocktail containing digoxin, metformin, furosemide, and rosuvastatin compared to rosuvastatin alone is most likely attributable to metformin, and only to a minor degree to furosemide. Reduced doses of metformin and furosemide are expected to eliminate mutual pharmacokinetic interactions in a drug transporter cocktail, which will be a valuable tool in drug development.
Acknowledgements
The authors thank Paul Tanswell for organizational and medical writing support during the preparation of this article. We also would like to thank the team of the Human Pharmacology Centre, Boehringer Ingelheim’s phase I unit for highly professional clinical conduct of the study. The authors acknowledge Gabriele Götz for data management, Patrik Faber and Renate Mang as well as SGS Cephac Europe for excellent bioanalytical work, Sven Schmidt for excellent support during the pharmacokinetic analysis of the trial, and Edith Hauel for medical writing of the clinical trial report. We would also like to thank Florian Gantner, head of the Translational Medicine and Clinical Pharmacology department, for his continued support.
Compliance with Ethical Standards
Funding
The study was funded by Boehringer Ingelheim Pharma GmbH & Co. KG, the Sponsor of the trial.
Conflicts of interest
Kathrin Hohl was contracted by Boehringer Ingelheim as an external statistician. All other authors are employees of Boehringer Ingelheim.
Ethical approval
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The clinical trial protocol was approved by the Ethics Commission of the State Chamber of Physicians of Baden-Württemberg, Stuttgart, Germany, and the Federal Institute for Drugs and Medicinal Products (BfArM), Bonn, Germany.
Informed consent
Written informed consent was obtained from all individual participants included in this trial.
Footnotes
Peter Stopfer and Thomas Giessmann contributed equally to this article.
References
- 1.König J, Müller F, Fromm MF. Transporters and drug-drug interactions: important determinants of drug disposition and effects. Pharmacol Rev. 2013;65(3):944–966. doi: 10.1124/pr.113.007518. [DOI] [PubMed] [Google Scholar]
- 2.Giacomini KM, Huang SM, Tweedie DJ, Benet LZ, Brouwer KL, Chu X, et al. Membrane transporters in drug development. Nat Rev Drug Discov. 2010;9(3):215–236. doi: 10.1038/nrd3028. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Zolk O, Fromm MF. Transporter-mediated drug uptake and efflux: important determinants of adverse drug reactions. Clin Pharmacol Ther. 2011;89(6):798–805. doi: 10.1038/clpt.2010.354. [DOI] [PubMed] [Google Scholar]
- 4.EMA-CHMP. Guideline on the investigation of drug interactions: final (CPMP/EWP/560/95/Rev. 1 corr. 2). 2012 Jun. http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2012/07/WC500129606.pdf. Accessed 13 March 2017.
- 5.US-FDA. Guidance for industry: drug interaction studies - study design, data analysis, implications for dosing, and labeling recommendations (draft guidance). 2012 Feb. http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm292362.pdf. Accessed 13 March 2017.
- 6.PMDA. Pharmaceuticals & Medical Device Agency-Japan. Drug interaction guideline for drug development and labeling recommendations (draft for public comment) 2014. http://www.solvobiotech.com/documents/Japanese_DDI_guideline_(draft)_2014Jan.pdf. Accessed 13 March 2017.
- 7.Streetman DS, Bleakley JF, Kim JS, Nafziger AN, Leeder JS, Gaedigk A, et al. Combined phenotypic assessment of CYP1A2, CYP2C19, CYP2D6, CYP3A, N-acetyltransferase-2, and xanthine oxidase with the “Cooperstown cocktail”. Clin Pharmacol Ther. 2000;68(4):375–383. doi: 10.1067/mcp.2000.109519. [DOI] [PubMed] [Google Scholar]
- 8.Turpault S, Brian W, Van Horn R, Santoni A, Poitiers F, Donazzolo Y, et al. Pharmacokinetic assessment of a five-probe cocktail for CYPs 1A2, 2C9, 2C19, 2D6 and 3A. Br J Clin Pharmacol. 2009;68(6):928–935. doi: 10.1111/j.1365-2125.2009.03548.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Stopfer P, Giessmann T, Hohl K, Sharma A, Ishiguro N, Taub ME, et al. Pharmacokinetic evaluation of a drug transporter cocktail consisting of digoxin, furosemide, metformin, and rosuvastatin. Clin Pharmacol Ther. 2016;100(3):259–267. doi: 10.1002/cpt.406. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Prueksaritanont T, Tatosian DA, Chu X, Railkar R, Evers R, Chavez-Eng C, et al. Validation of a microdose probe drug cocktail for clinical drug interaction assessments for drug transporters and CYP3A. Clin Pharmacol Ther. 2016 doi: 10.1002/cpt.525. [DOI] [PubMed] [Google Scholar]
- 11.Ebner T, Ishiguro N, Taub ME. The use of transporter probe drug cocktails for the assessment of transporter-based drug–drug interactions in a clinical setting—proposal of a four component transporter cocktail. J Pharm Sci. 2015;104(9):3220–3228. doi: 10.1002/jps.24489. [DOI] [PubMed] [Google Scholar]
- 12.Lee D, Roh H, Son H, Jang SB, Lee S, Nam SY, et al. Pharmacokinetic interaction between rosuvastatin and metformin in healthy Korean male volunteers: a randomized, open-label, 3-period, crossover, multiple-dose study. Clin Ther. 2014;36(8):1171–1181. doi: 10.1016/j.clinthera.2014.06.004. [DOI] [PubMed] [Google Scholar]
- 13.US-FDA. CRESTOR (rosuvastatin calcium) tablets, Prescribing Information 2010. https://www.accessdata.fda.gov/drugsatfda_docs/label/2010/021366s016lbl.pdf. Accessed 20 June 2017.
- 14.US-FDA. Guidance for industry: bioanalytical method validation. 2001 May. http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm070107.pdf. Accessed 13 March 2017.
- 15.EMA-CHMP. Guideline on bioanalytical method validation EMEA/CHMP/EWP/192217/200921. 2011 Jul. http://www.ema.europa.eu/ema/pages/includes/document/open_document.jsp?webContentId=WC500109686. Accessed 13 March 2017.
- 16.Kupper LL, Hafner KB. How appropriate are popular sample size formulas? Am Stat. 1989;43:101–105. [Google Scholar]
- 17.Prueksaritanont T, Chu X, Evers R, Klopfer SO, Caro L, Kothare PA, et al. Pitavastatin is a more sensitive and selective organic anion-transporting polypeptide 1B clinical probe than rosuvastatin. Br J Clin Pharmacol. 2014;78(3):587–598. doi: 10.1111/bcp.12377. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kitamura S, Maeda K, Wang Y, Sugiyama Y. Involvement of multiple transporters in the hepatobiliary transport of rosuvastatin. Drug Metab Dispos Biol Fate Chem. 2008;36(10):2014–2023. doi: 10.1124/dmd.108.021410. [DOI] [PubMed] [Google Scholar]
- 19.Abe K, Bridges AS, Yue W, Brouwer KL. In vitro biliary clearance of angiotensin II receptor blockers and 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors in sandwich-cultured rat hepatocytes: comparison with in vivo biliary clearance. J Pharmacol Exp Ther. 2008;326(3):983–990. doi: 10.1124/jpet.108.138073. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Elsby R, Hilgendorf C, Fenner K. Understanding the critical disposition pathways of statins to assess drug–drug interaction risk during drug development: it’s not just about OATP1B1. Clin Pharmacol Ther. 2012;92(5):584–598. doi: 10.1038/clpt.2012.163. [DOI] [PubMed] [Google Scholar]
- 21.Bosgra S, van de Steeg E, Vlaming ML, Verhoeckx KC, Huisman MT, Verwei M, et al. Predicting carrier-mediated hepatic disposition of rosuvastatin in man by scaling from individual transfected cell-lines in vitro using absolute transporter protein quantification and PBPK modeling. Eur J Pharm Sci. 2014;65:156–166. doi: 10.1016/j.ejps.2014.09.007. [DOI] [PubMed] [Google Scholar]
- 22.Ho RH, Tirona RG, Leake BF, Glaeser H, Lee W, Lemke CJ, et al. Drug and bile acid transporters in rosuvastatin hepatic uptake: function, expression, and pharmacogenetics. Gastroenterology. 2006;130(6):1793–1806. doi: 10.1053/j.gastro.2006.02.034. [DOI] [PubMed] [Google Scholar]
- 23.Lou XY, Zhang W, Wang G, Hu DL, Guo D, Tan ZR, et al. The effect of Na+/taurocholate cotransporting polypeptide (NTCP) c.800C>T polymorphism on rosuvastatin pharmacokinetics in Chinese healthy males. Pharmazie. 2014;69(10):775–779. [PubMed] [Google Scholar]
- 24.Kusuhara H, Ito S, Kumagai Y, Jiang M, Shiroshita T, Moriyama Y, et al. Effects of a MATE protein inhibitor, pyrimethamine, on the renal elimination of metformin at oral microdose and at therapeutic dose in healthy subjects. Clin Pharmacol Ther. 2011;89(6):837–844. doi: 10.1038/clpt.2011.36. [DOI] [PubMed] [Google Scholar]
- 25.Graham GG, Punt J, Arora M, Day RO, Doogue MP, Duong JK, et al. Clinical pharmacokinetics of metformin. Clin Pharmacokinet. 2011;50(2):81–98. doi: 10.2165/11534750-000000000-00000. [DOI] [PubMed] [Google Scholar]
- 26.Ponto LL, Schoenwald RD. Furosemide (frusemide). A pharmacokinetic/pharmacodynamic review (Part I) Clin Pharmacokinet. 1990;18(5):381–408. doi: 10.2165/00003088-199018050-00004. [DOI] [PubMed] [Google Scholar]
- 27.Glucophage (metformin hydrochloride tablets) 500 mg, 850 mg, and 1000 mg, Rx only (product information). Physician’s Desk Ref. 2001;1005–9.