Abstract
AIM
To investigate whether oseltamivir enhances the anticoagulant effect of warfarin and to evaluate any pharmacokinetic (PK) interaction between the agents.
METHODS
Twenty volunteers (mean age 62 years) receiving daily warfarin and with INR values of 2.0–3.5 during the previous 2 weeks were randomized to concomitant oseltamivir 75 mg twice daily for 4.5 days or warfarin alone in a two-way cross-over design with a 4–8 day wash-out. Anticoagulant effects were assessed by calculating overall [AUEC(0,96 h)] and observed maximum effect (Emax) increase from baseline in INR, decrease from baseline in factor VIIa, and change in vitamin K1 concentrations. Plasma pharmacokinetics of (R)- and (S)-warfarin and oseltamivir were also assessed.
RESULTS
For both treatments, changes in INR and factor VIIa during treatment were small; for net AUEC(0,96 h), least square mean values were −9.53 (oseltamivir + warfarin) and −1.69 h (warfarin alone) for INR (difference −7.84 h, 90% CI −18.86, 3.17 h), and 1.56 and 0.54 kIU l−1 h, respectively, for factor VIIa (difference, 1.01 kIU l−1 h; 90% CI −1.18, 3.21). Differences between the treatments in Emax increase from baseline for INR, decrease from baseline for factor VIIa and change from baseline in vitamin K1 concentration were not statistically significant. Oseltamivir did not alter warfarin pharmacokinetics. Oseltamivir was well tolerated in this study with no clinically significant adverse safety findings.
CONCLUSION
Concomitant administration of oseltamivir for 4.5 days to volunteers on daily warfarin had little or no effect on warfarin pharmacokinetics and no effect on pharmacodynamics.
Keywords: coagulation, interaction, oseltamivir, pharmacodynamics, pharmacokinetics, warfarin
WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT
Although oseltamivir had not previously been reported to cause haematological toxicity, the European Medicines Agency (EMEA) asked Roche to investigate potential interactions between oseltamivir and warfarin after isolated reports of an enhancement of the anticoagulant effect of warfarin during oseltamivir treatment.
WHAT THIS STUDY ADDS
In individuals taking warfarin at stable dosages for chronic vascular or cardiac conditions, there is no evidence for a pharmacodynamic or pharmacokinetic interaction between oseltamivir and warfarin that could alter the anticoagulant effect of the latter.
Introduction
The neuraminidase inhibitor oseltamivir (Tamiflu®; F Hoffmann-La Roche Ltd) has become established as an effective agent for the treatment and prophylaxis of influenza in adults, including those over the age of 65 years, and children aged older than 1 year. Experience from both randomized controlled trials and clinical practice has shown that the drug is well tolerated, with nausea and vomiting being the most common adverse event in the target population [1]. Furthermore, there are no data to suggest that the drug may be associated with haematological adverse reactions, and pharmacology studies have shown low levels of plasma protein binding and metabolism independent of the cytochrome P (CYP) 450 and glucuronidase systems which suggests that clinical drug interactions affecting coagulation processes are unlikely [2, 3].
During the period between 1 January 1999 and 31 October 2005, the Canadian health authorities received reports of 19 people in whom oseltamivir had been associated with an enhancement of the anticoagulant effect of warfarin. The effect was seen between 1 and 11 days after the start of the 5 day oseltamivir treatment period [4]. Increases in INR ranged from 3.2 to 10.9 and six patients required vitamin K treatment, although causality remains unclear. In three cases, the INR increases were noted after an increase in warfarin dosage following the introduction of oseltamivir, and in three other cases decreases in INR were observed during oseltamivir therapy in the absence of a change in warfarin dosage. Cases of bleeding in warfarin-treated patients who received oseltamivir in Europe were also reported. In response to this, the European Medicines Agency (EMEA) requested that Roche conduct a clinical pharmacology study to investigate potential interactions between oseltamivir and warfarin. The primary objective of this study was to determine whether oseltamivir has any affect on the pharmacodynamics of warfarin in volunteers receiving stable dosages of the drug to treat conditions such as thrombosis.
Methods
Study design
The study employed a two-way cross-over design, and treatments were administered open-label (with random allocation to one of two treatment sequences). The study was conducted at a clinical trials unit in the UK, and was conducted in compliance with the principles of the Declaration of Helsinki and in accordance with Good Clinical Practice (GCP). The protocol and all amendments were approved by the institutional review board for the study centre and all participants gave written informed consent.
Study measures
Pharmacodynamic variables measured to assess whether oseltamivir had any affect on the pharmacodynamics of warfarin included the International Normalized Ratio (INR) and the concentration of factor VIIa. A rise in the former and/or a fall in the latter would signify enhancement of warfarin's anticoagulant effect. Changes in vitamin K1 concentration were also assessed. The effect, if any, of oseltamivir on warfarin pharmacokinetics (and vice versa), and the safety and tolerability of both drugs when given together, were also studied.
Patient selection
Men or women aged 18–85 years and with a body mass index (BMI) of 18–32 kg m−2 were eligible for study entry if they were receiving warfarin once daily for at least 4 weeks before screening and their coagulation had been regularly monitored to assess stability of the INR value. They also had to have a stable INR value in the range 2.0–3.5 in the 2 weeks before starting study treatment, assessed through daily self-monitoring during this period. Any participant with a creatinine clearance of <60 ml min−1 (by Cockcroft-Gault formula) or whose warfarin dose had changed between the screening visit and the day before starting study treatment was excluded from the study, as were those with a history of coagulation disorders and those who were taking any of the following: vitamin K in any form; any drug or agent known to induce or inhibit CYP450 enzymes or any other non-chronically administered drug known to interact either with warfarin pharmacokinetics, INR, vitamin K or endogenous factor VIIa levels.
Interventions
The study design consisted of two 5-day treatment periods, periods 1 and 2, separated by a wash-out period of 4–8 days. For period 1, the volunteers were randomly allocated to either treatment A (warfarin alone, 10 volunteers) or treatment B (warfarin plus oseltamivir, 10 volunteers). For period 2, volunteers received whichever treatment they did not receive in period 1, completing the cross-over sequence.
The dosages of the two study treatments were as follows. Warfarin was taken orally once daily, on the same dose regimen as before study entry. This meant that the warfarin dosage and presentation was different for each volunteer (titrated as part of their usual clinical care to target a common pharmacodynamic outcome/effect in each). The oseltamivir dosage for all volunteers was 75 mg twice daily for 4 days plus a single dose on the morning of day 5. In the warfarin-only treatment period, no oseltamivir placebo was used. The nine-dose oseltamivir regimen (compared with ten in usual clinical practice) was used for practical purposes, to measure AUC(0,24 h) for oseltamivir at steady-state kinetics (which are known to be attained by day 4 of dosing) and since the drug was not being given for therapeutic reasons.
Assessment schedule
Blood was taken for safety analysis (haematology and biochemistry) and urine collected for urinalysis on each of the 5 days of each treatment period (immediately before oseltamivir dosing), on the days before and after each treatment period and on the day of the follow-up visit (between 4 and 12 days after the end of period 2). Levels of two of the analytes measured for PD outcomes (INR and factor VIIa) were assayed using these safety blood samples (for the 5 days of the treatment period only). Vitamin K1 concentrations were measured on day 1 of each treatment period and 24 h post-dose on day 6 after each treatment period.
Blood was taken for pharmacokinetic analysis of warfarin on either days 4 and 5 or days 5 and 6 of treatment (see below) and for oseltamivir, it was taken on days 1, 5 and 6. The sampling schedule for warfarin was at 0 h (pre-dose) then 1, 2, 4, 8, 12, and 24 h post-dose; if warfarin was given in the evening, the 0 h sample was in the evening of day 4, and if it was given in the morning, the 0 h sample was in the morning of day 5. The schedule for oseltamivir and oseltamivir carboxylate (in the warfarin + oseltamivir treatment period) was 0 h (pre-dose) and 0.5, 1, 1.5, 2, 3, 4, 6, 8, 10 and 12 h post-dose on day 1, and at the same times on day 5 with two extra samples at 18 and 24 h post-dose.
Vital signs and ECG were assessed on all days of the treatment period, on the 3 days before each treatment period, the day after each treatment period and at the follow-up visit. The same timetable applied to patients' self-assessment of INR levels using a Coaguchek® device and test strips (Roche Diagnostics GmbH, Mannheim, Germany), except that readings were also taken on each of the 11 days before the first 3 day pre-treatment period. Patients were trained in how to use this device if necessary, and readings were entered by the patients on an INR record card.
Assays
Total and unbound concentrations of (R)- and (S)-warfarin were analysed by TurboIonSpray liquid chromatography/tandem mass spectrometry (Applied Biosystems, Langen, Germany), using column switching LC/MS/MS in the negative ion mode. The lower limits of quantification (LLOQs) were 1 ng ml−1 for total (R)- and (S)-warfarin and 0.1 ng ml−1 for free (R)- and (S)-warfarin; the calibration ranges were up to 2500 ng ml−1 for total warfarin and up to 10 ng ml−1 for the unbound form. The assays were accurate and reproducible. The inter-assay precision (coefficient of variation) of quality control samples ranged between 5.0–7.2% for total (R)-warfarin, between 4.7–7.7% for total (S)-warfarin, between 3.0–7.5% for free (R)-warfarin, and between 2.5–6.4% for free (S)-warfarin. There was no marked inaccuracy in the results from these quality control (QC) samples: mean accuracy 92.4% (n = 21) to 94.3% (n = 21) for total (R)-warfarin, 93.7% (n = 21) to 98.7% (n = 21) for total (S)-warfarin, 95.7% (n = 12) to 104.0% (n = 12) for free (R)-warfarin, and 97.0% (n = 15) to 103.8% (n = 15) for free (S)-warfarin.
Oseltamivir and its carboxylate metabolite were assayed using a validated LC/MS/MS method. Quantification was achieved using deuterated internal standards. Detection was by tandem mass spectrometry, monitoring positive ions produced in the TurboIonSpray source of a PE Sciex API 3000 mass spectrometer (Applied Biosystems, Langen, Germany). The LLOQs for a 100 µl sample were 1 ng ml−1 for oseltamivir and 10 ng ml−1 for the carboxylate moiety and the calibration ranges were up to 250 ng ml−1 for oseltamivir and 10 000 ng ml−1 for the carboxylate. The performance of the assay for all study samples was satisfactory throughout the study. The inter-assay precision (percentage coefficient of variation) of the QC samples ranged between 2.9–8.5% for oseltamivir (prodrug) and 4.0–7.9% for the active metabolite. There was no marked inaccuracy in the results from these QC samples: mean accuracy 96.0% (n = 26) to 104.7% (n = 26), prodrug and mean accuracy 94.4% (n = 26) to 102.3% (n = 26), active metabolite.
Study endpoints
Pharmacodynamics (PD)
For two of the three PD variables (INR and factor VIIa activity), parameters describing changes over the treatment period were derived by non-compartmental analysis (venous INR values were used in the pharmacodynamic analyses). The absolute change in INR and factor VIIa from baseline was also assessed and where baseline was defined as day 1 pre-dose of each treatment period. For INR value and factor VIIa activity, three pharmacodynamic parameters were derived, especially to give a measure of the overall effect during oseltamivir dosing. The net area under the plasma effect–time curve over 96 h (AUEC(0,96 h)) was calculated using the linear trapezoidal rule; this was the area under the effect–time curve and above the baseline minus the area above the curve and below the baseline during the 5-day period. The maximum observed change from baseline (Emax) and the time at which this value was reached (tmax) were also calculated, although it should be noted that for each analyte, only one direction of effect was of interest to assess the potential bleeding risk, i.e. the largest increase from baseline in INR and the largest fall from baseline in factor VIIa. For the statistical analysis the uncorrected Emax values were used.
For the third PD variable, vitamin K1, the absolute change in concentration from pre-dose concentration on day 1 to day 6 was calculated.
Pharmacokinetics (PK)
PK measures for both warfarin and oseltamivir were also derived using non-compartmental analysis. PK parameters were derived for each of the warfarin isomers (R and S) and for both total plasma warfarin and the free (non-protein bound) plasma fraction. The parameters derived were AUC(0,12 h), AUC(0,24 h), Cmax, tmax, t1/2 and CL/F (plasma clearance after oral dosing). AUC and its derived parameters were calculated using the linear trapezoidal rule. For oseltamivir and its carboxylate metabolite, the following PK parameters were derived from day 1 (single dose) and day 5 (steady-state) data: AUC(0,12 h), Cmax, tmax, and t1/2. CL/F was also derived for oseltamivir free base only. Additionally, AUC(0,24 h) was derived on day 5 only (for both moieties).
Because volunteers were taking different daily doses of warfarin to achieve a stable anticoagulant effect (INR value), AUC and Cmax values for each volunteer were normalized before analysis, each individual's value being divided by the average daily dose over the respective warfarin dosing regimen cycle (i.e. 1 week).
Tolerability and safety
Adverse events (AEs) were monitored throughout the study and during a follow-up period of between 4 and 12 days after the end of period 2. Safety was assessed by laboratory test results (haematology, biochemistry and urinalysis), vital signs measurement, ECG and physical examination at the follow-up visit. In addition, patients were asked to report all AEs prior to follow up.
Statistical analysis
For the three PD measures, a mixed linear model analysis was ‘applied’ to parameters derived from the non-compartmental analysis. To assess the effect of oseltamivir on the INR after multiple dosing, warfarin pharmacodynamic parameters following dosing with oseltamivir (Treatment B) were compared with those following dosing with warfarin alone (Treatment A). Point estimates (least squares means) and 90% confidence intervals were calculated using contrasts from the mixed model.
The statistical model:
 |
was used for the primary PD analyses, where yij = the PD parameter obtained for subject i (I = 1, …, n) on treatment τd(i,j) (Treatment A or B) in period j (j = 1 or 2).
In this model, τd(i,j), pj and τd(i,j)× pj are fixed effects associated with the treatment, period and the treatment by period interaction, si are random subject effects and εij denotes random error. The random variables si and εij are assumed to be independent and have normal distributions with means of zero and standard deviations σs and σε, respectively.
For the analysis of warfarin pharmacokinetics, least square means and 90% CI were derived from the linear mixed model for the parameters AUC(0,24 h) and Cmax, for both warfarin isomers (total drug in plasma and unbound drug).
Sample size
A total of 20 volunteers were enrolled (10 to each treatment sequence). If nine volunteers are included in each treatment sequence there will be at least 80% power for AUC, and at least 90% power for Cmax, to show that the 90% CI for the ratio of mean values, oseltamivir plus warfarin compared with warfarin alone (log scale) will be within 0.8 to 1.25. This is assuming that the expected ratio of means is 1.0 and the cross-over root mean square errors are 0.20 and 0.17 for AUC and Cmax, respectively (log scale). These same calculations for 80% power apply for INR and factor VII assuming the cross-over root mean square error is 0.20 (log scale).
Results
Disposition and baseline clinical characteristics
As no patients withdrew during the study, data from 20 volunteers were evaluable for PD, PK, tolerability and safety analyses. The mean values at screening for the age, sex, weight, BMI and INR of patients are shown in Table 1. The average daily dose of warfarin taken by each patient ranged from 0.79 to 7.71 mg. The within-subject variability in INR for some individual patients prior to treatment was noteworthy, as shown by the venous INR results at screening and prior to first dosing of the study drug (Figure 1).
Table 1.
Demographics and clinical characteristics at screening. Only INR was measured at screening
| Characteristic | Value (mean and SD) |
|---|---|
| Men/women (n) | 14/6 |
| Age (years) | 62.4 (8.8) |
| Weight (kg) | 82.2 (14.3) |
| BMI (kg m−2) | 28.1 (2.32) |
| Analyte | Mean (range) |
|---|---|
| INR (venous), ratio | 2.2 (1.7–2.8) |
Factor VIIa and vitamin K were not measured. BMI, body mass index; INR, international normalized ratio.
Figure 1.
Change in individual venous INR values prior to enrolment, includes values at screening and prior to first dosing of study drug. INR, international normalized ratio; Pre-dose, period 1 (day 1) pre-dose INR value; Screening was performed between 43 and 15 days before the subject's first dosing day. Note: Each line represents an individual subject; each subject is identified in the legend on the right hand side of the plot. 1001 (
); 1002 (
); 1003 (
); 1004 (
); 1005 (
); 1006 (
); 1007 (
); 1008 (
); 1009 (
); 1010 (
); 1011 (
); 1012 (
); 1013 (
); 1014 (
); 1015 (
); 1016 (
); 1017 (
); 1018 (
); 1019 (
); 1020 (
)
All patients had at least one existing disease, the most common of which were vascular diseases (n = 14), typically deep venous thrombosis or hypertension (n = 7 for each). Six patients had pulmonary embolism and four had atrial fibrillation. After warfarin (taken by all patients), the most common co-prescribed medications were statins (n = 12), ACE inhibitors and/or angiotensin receptor blockers (n = 9), and other antihypertensive drugs such as β-adrenoceptor blockers and calcium channel blockers (n = 8).
Pharmacodynamics
The changes from baseline in INR values over the treatment period were small, and similar for the two study treatments: in the warfarin-only group, the mean value of net AUEC(0,96 h) was −2.16 h and in the warfarin + oseltamivir group, −9.06 h. The corresponding mean values for Emax (maximum observed increase from baseline) were 0.3 and 0.1, respectively (Table 2). In both groups, the time at which the maximum increase from baseline in INR value occurred (tmax) varied considerably between individuals, with medians of 24 h for warfarin alone and 0 h for warfarin + oseltamivir. The changes from baseline in INR over the treatment period in each individual were quite similar, except for two patients receiving warfarin alone, who had substantially higher increases from baseline in INR values on some days, and one patient receiving warfarin + oseltamivir who had a sustained reduction from baseline in INR. In all three cases, these data points lay outside the range seen for other 17 patients in this study (Figure 2).
Table 2.
Changes during the treatment period in INR (n = 20), factor VIIa (n = 19) and vitamin K1 (n = 20)
| Parameter value (mean and range) | ||
|---|---|---|
| Analyte and parameter | Warfarin alone | Warfarin + oseltamivir |
| INR | ||
| Baseline | 2.2 (1.6–3.0) | 2.1 (1.5–3.3) |
| Net AUEC(0,96 h) (h) | −2.16 (−27.6–78.0) | −9.06 (−93.6–18.0) |
| Emax, absolute value | 2.4 (1.6–4.6) | 2.2 (1.6–3.5) |
| Emax, change from baseline | 0.3 (0.0–2.1) | 0.1 (0.0–0.5) |
| tmax (median and range) (h) | 24 (0–96) | 0 (0–96) |
| Factor VIIa activity | ||
| Baseline (kIU l−1) | 0.352 (0.17–0.68) | 0.363 (0.21–0.60) |
| Net AUEC(0,96 h) (kIU l−1 h) | 0.568 (−9.48–7.56) | 1.45 (−5.16–12.5) |
| Emax, absolute value (kIU l−1) | 0.301 (0.150–0.590) | 0.320 (0.200–0.560) |
| Emax, change from baseline (kIU l−1) | −0.0505 (−0.21–0.00) | −0.0432 (−0.15–0.00) |
| tmax (median and range) (h) | 96 (0–96) | 72 (0–96) |
| Vitamin K1 | ||
| Baseline (ng l−1) | 571 (168–1410) | 562 (217–2040) |
| Day 6, absolute value (ng l−1) | 876 (286–1900) | 833 (296–1730) |
| Change from baseline at day 6 (ng l−1) | 305 (−371–1060) | 271 (−1100–1310) |
For definition of net as applied to AUEC values, see text. AUEC(0,96 h), area under the effect curve up to 96 h after dosing; Baseline, predose value on day 1 of each treatment period; Emax, maximum observed increase (INR) or decrease (factor VIIa); INR, international normalized ratio; tmax, time of maximum change from baseline.
Figure 2.
Changes from baseline in venous INR values over the 5-day treatment period for each volunteer and mean change for warfarin alone and warfarin plus oseltamivir. Treatment A: warfarin alone o.d. p.o. (subject's usual dosing regimen). Treatment B: warfarin o.d. p.o. + oseltamivir 75 mg twice daily p.o. for 4 days with a single dose on day 5. Note: overlaid mean plots are provided for both treatments: Treatment A (thick black line with circular markers) and treatment B (thick grey line with triangular markers). o.d., once daily; p.o., orally; b.i.d., twice daily
Changes from baseline in factor VIIa activity were also slight for both treatments, as shown by the mean values of AUEC(0,96 h), Emax and tmax in Table 2. For both treatments, the trend was for individual factor VIIa concentrations to increase from baseline on day 1, but then to fall slowly for the rest of the treatment period (Figure 3). The largest individual falls from baseline values were seen on day 5 (0.21 kIU l−1 for warfarin alone and 0.15 kIU l−1 for warfarin + oseltamivir), although the median tmax was 96 h for warfarin alone and 72 h for warfarin + oseltamivir.
Figure 3.
Changes from baseline in factor VIIa activity over the 5-day treatment period for each volunteer and mean change for warfarin alone and warfarin plus oseltamivir. Treatment A: warfarin alone o.d. p.o. (subject's usual dosing regimen). Treatment B: warfarin o.d. p.o. + oseltamivir 75 mg b.i.d. p.o. for 4 days with a single dose on day 5. Note: overlaid mean plots are provided for both treatments: Treatment A (thick black line with circular markers) and Treatment B (thick grey line with triangular markers). o.d., once daily; p.o., orally; b.i.d., twice daily
The mean change in vitamin K1 concentrations between baseline (day 1) and 24 h post oseltamivir dose on day 5 was similar for the two treatments, as shown in Table 2. The individual changes are shown in Figure 4.
Figure 4.
Vitamin K1 concentrations at baseline and on day 5 for each volunteer and mean values for warfarin alone and warfarin plus oseltamivir. Treatment A: warfarin alone o.d. p.o. (subject's usual dosing regimen). Treatment B: warfarin o.d. p.o. + oseltamivir 75 mg b.i.d. p.o. for 4 days with a single dose on day 5. Note: overlaid mean plots are provided for both treatments: Treatment A (thick black line with circular markers) and Treatment B (thick grey line with triangular markers). o.d., once daily; p.o., orally; b.i.d., twice daily
Comparisons of PD parameters for the two study treatments, expressed as either the least square means for the difference in AUEC(0,96 h) and vitamin K1 change from baseline or the least square means for the Emax ratios, are shown in Table 3. The 90% CI for the differences and ratios do not suggest any significant differences between treatments for net AUEC of INR and factor VIIa, and vitamin K1 change from baseline values. Confidence limits for the ratios of mean Emax INR and factor VIIa activity fell within the limits of 0.8 and 1.25 supporting a lack of pharmacodynamic interaction between warfarin and oseltamivir. The mixed model analysis showed a statistically significant effect of treatment period on the change from baseline in vitamin K1 (P < 0.05), indicating a possible difference in mean vitamin K1 values between treatment periods.
Table 3.
Least square mean estimates for three PD measures, comparing the two study treatments by either differences or ratios
| Parameter value (least square mean) | Arithmetic difference* (90% confidence intervals) | ||
|---|---|---|---|
| Analyte and parameter | Warfarin alone | Warfarin + oseltamivir | |
| INR | |||
| Net AUEC(0,96 h) (h) | −1.69 | −9.53 | −7.84 (−18.9, 3.17) |
| Emax, absolute value | 2.32 | 2.20 | 0.95 (0.90, 1.00) |
| Factor VIIa activity | |||
| Net AUEC(0,96 h) (kIU l−1 h) | 0.54 | 1.56 | 1.01 (−1.18, 3.21) |
| Emax, absolute value (kIU l−1) | 0.29 | 0.30 | 1.05 (0.97, 1.13) |
| Vitamin K1 | |||
| Change from baseline at day 6 (ng l−1) | 307 | 268 | −38.5 (−129, 52.8) |
For definition of net as applied to AUEC values, see text.
Ratio for Emax. AUEC(0,96 h), area under the effect curve up to 96 h after dosing; Emax, maximum observed effect in INR or Factor VIIa activity; INR, international normalized ratio.
Pharmacokinetics
Warfarin
Mean steady-state values for the primary warfarin PK parameters AUC(0,24 h) and Cmax were very similar for the two study treatments. This applied to both warfarin isomers in their total and unbound forms. The same degree of similarity between treatments was seen in the estimates of all secondary parameters.
Mean values for primary and secondary PK parameters are shown in Table 4. The results of least square means analysis confirmed a high degree of similarity in warfarin PK between study treatments (Table 5). The 90% CIs for the ratio of one treatment to the other fell within the limits of 0.8 to 1.25 except one that marginally fell outside (Cmax for total (S)-warfarin: 90% CI 0.78, 0.99) supporting a lack of a clinically relevant pharmacokinetic interaction between warfarin and oseltamivir. Figure 5 illustrates the similarity between treatments in plasma concentration vs. time curves for bound and free (R)- and (S)-warfarin.
Table 4.
Pharmacokinetic parameters for total and unbound warfarin (R and S isomers) for warfarin only (W only) and warfarin plus oseltamivir (W + O) (n = 20 for each unless indicated)
| (S)-warfarin | (R)-warfarin | |||||||
|---|---|---|---|---|---|---|---|---|
| Total | Unbound | Total | Unbound | |||||
| W only | W + O | W only | W + O | W only | W + O | W only | W + O | |
| Primary parameters: | ||||||||
| Cmax/Davg (ng ml−1 mg−1) | 394 (259) | 330 (197) | 1.87 (1.3) | 1.68 (1.13) | 436 (243) | 375 (96.3) | 2.70 (1.45) | 2.37 (0.60) |
| AUC(0,24 h)/Davg (ng ml−1 h mg−1) | 6770 (4500) | 6460 (4310) | 35.2 (26.1) | 33.9 (24.1) | 7790 (2480) | 7360 (1960) | 49.4 (13.4) | 48.3 (11.9) |
| Secondary parameters: | ||||||||
| AUC(0,12 h)/Davg (ng ml−1 h mg−1) | 3610 (2200) | 3380 (2100) | 18.6 (13.0) | 17.8 (12.3) | 4140 (1320) | 3870 (1020) | 25.9 (7.84) | 24.8 (6.36) |
| tmax (h)* | 2.00 (1.00–8.13) | 4.00 (0.00–23.92) | 3.01 (1.00–8.00) | 2.00 (1.00–11.83) | 3.00 (1.00–11.83) | 4.00 (1.00–23.92) | 4.00 (1.00–11.83) | 3.00 (1.00–11.83) |
| t1/2 (h) | 45.4 (23.0)† | 45.6 (11.7)‡ | [n.c.] | [n.c.] | 56.1 (26.1)§ | 53.6 (12.2)§ | [n.c.] | [n.c.] |
| CL/F (l h−1) | 0.198 (0.0895) | 0.208 (0.0959) | 38.5 (17) | 38.2 (15.1) | 0.139 (0.0361) | 0.145 (0.0374) | 21.7 (5.74) | 22.0 (5.54) |
Values are from day 5, and are mean and SD unless stated otherwise.
Median and range
n = 13
n = 11
n = 12.
Abbreviations: AUEC(0,24 h), area under the effect curve up to 24 h after dosing; AUEC(0,12 h), area under the effect curve up to 12 h after dosing; CL/F, total body clearance; Cmax, maximum observed concentration; Davg, individual average dose; n.c., not calculated (because n < 10); tmax, time of maximum observed concentration; t1/2, terminal elimination half-life.
Table 5.
Least square mean estimates for each warfarin PK measure on day 5, comparing the two study treatments by ratios
| Parameter value (least square mean) | |||
|---|---|---|---|
| Analyte and parameter | Warfarin alone | Warfarin + oseltamivir | Ratio (90% confidence intervals) |
| (S)-warfarin, total | |||
| AUC(0,24 h) (ng ml−1 h) | 10 467 | 9996 | 0.96 (0.90, 1.01) |
| Cmax (ng ml−1) | 603 | 529 | 0.88 (0.78, 0.99) |
| (S)-warfarin, unbound | |||
| AUC(0,24 h) (ng ml−1 h) | 87.17 | 85.72 | 0.98 (0.95, 1.02) |
| Cmax (ng ml−1) | 4.56 | 4.21 | 0.92 (0.85, 1.01) |
| (R)-warfarin, total | |||
| AUC(0,24 h) (ng ml−1 h) | 13 662 | 13 006 | 0.95 (0.90, 1.01) |
| Cmax (ng ml−1) | 728 | 664 | 0.91 (0.83, 1.00) |
| (R)-warfarin, unbound | |||
| AUC(0,24 h) (ng ml−1 h) | 87.17 | 85.72 | 0.98 (0.95, 1.02) |
| Cmax (ng ml−1) | 4.56 | 4.21 | 0.92 (0.85, 1.01) |
AUC(0,24 h), area under the concentration–time curve up to 24 h after dosing; Cmax, maximum concentration
Figure 5.
Normalized mean plasma concentration vs. time profiles for free and bound (R)- and (S)-warfarin at steady state according to treatment. Treatment A: warfarin alone; treatment B: warfarin + oseltamivir. (A) Free warfarin. Treatment A: warfarin alone o.d. p.o. (subject's usual dosing regimen).Treatment B: warfarin o.d. p.o. + oseltamivir 75 mg b.i.d. p.o. for 4 days with a single dose on day 5. Normalized free R-warfarin, Treatment A (
); Normalized free R-warfarin, Treatment B (
); Normalized free S-warfarin, Treatment A (
); Normalized free S-warfarin, Treatment B (
). (B) Bound warfarin. Treatment A: warfarin alone o.d. p.o. (subject's usual dosing regimen). Treatment B: warfarin o.d. p.o. + oseltamivir 75 mg b.i.d. p.o. for 4 days with a single dose on day 5. Normalized total R-warfarin, Treatment A (
); Normalized total R-warfarin, Treatment B (
); Normalized total S-warfarin, Treatment A (
); Normalized total S-warfarin, Treatment B (
)
Oseltamivir
PK parameters on day 1 and day 5 in volunteers receiving oseltamivir + warfarin in the present study (Table 6) were compared with other published data [5]. Visual inspection suggested exposure to oseltamivir as shown by AUC(0,12 h) values to be comparable with that in healthy volunteers in previous studies.
Table 6.
Summary of mean (SD) pharmacokinetic data for oseltamivir and oseltamivir carboxylate derived from day 1 (single dose) and day 5 (steady-state)
| Day 1 (single dose) | Day 5 (steady-state) | |||
|---|---|---|---|---|
| Oseltamivir (n = 20) | Oseltamivir carboxylate (n = 20) | Oseltamivir (n = 20) | Oseltamivir carboxylate (n = 20) | |
| Primary parameters: | ||||
| Cmax (ng ml−1) | 88.9 (50.0) | 367 (107) | 91.2 (47.2) | 571 (160) |
| AUC(0,12 h) (ng ml−1 h) | 141 (38.3) | 2990 (793) | 169 (36.3) | 5110 (1630) |
| Secondary parameters: | ||||
| AUC(0,24 h) (ng ml−1 h) | – | – | 176 (38.5) | 7180 (2550) |
| tmax (h)* | 0.50 (0.50–3.00) | 4.00 (2.00–6.00) | 0.75 (0.50–3.00) | 4.00 (3.00–6.00) |
| t1/2 (h)** | 1.77 (0.686) | 6.17 (2.07) | 4.00 (3.08) | 8.19 (1.90) |
| CL/F (l h−1) | 558 (134) | – | 463 (104) | – |
Values are mean and SD unless stated otherwise. Oseltamivir was only administered (and PK samples taken) in one treatment period (treatment B).
Median and range.
t1/2 is ‘apparent’ for oseltamivir carboxylate.
Abbreviations: AUEC(0,12 h), area under the effect curve up to 12 h after dosing; AUEC(0,24 h), area under the effect curve up to 24 h after dosing; CL/F, total body clearance; Cmax, maximum observed concentration; tmax, time of maximum observed concentration; t1/2, terminal elimination half-life.
Tolerability
Eleven AEs were reported (five during treatment A and six during treatment B), of which eight were judged not to be related to study treatment. Two reports of headache (one of moderate intensity and one mild) were judged to be possibly related to study treatment, and one report of mild diarrhoea was judged to be remotely related to study treatment. One serious AE, angina pectoris, was recorded in a male volunteer 2 days after the end of period 2 (during which he received treatment B). The volunteer had co-existing cardiovascular disease (coronary artery disease, atrial fibrillation and hypertension). The event, which resolved the following day after vasodilator treatment, was judged to be unrelated to study treatment, and there were no clinically relevant changes in the patient's coagulation data between baseline (day 1 predose of period 2) and the day of event.
Safety
No clinically meaningful changes were seen in any laboratory test value during either study treatment, except for one episode of mild hypoglycaemia in a volunteer with type 2 diabetes who received treatment B during period 2. The event was judged unrelated to study treatment and resolved quickly after treatment. Isolated laboratory test results that were found to be out of normal range were judged to be clinically unremarkable in view of the age and medical history of the volunteers affected.
No clinically meaningful changes were seen in other safety assessments (ECG and vital signs) during the study.
Discussion
In the current study, concomitant administration of oseltamivir for 4.5 days to volunteers receiving daily warfarin therapy with a stable coagulation status (INR of 2.0–3.5) had no clinically relevant effect on the anticoagulant effect of warfarin. Oseltamivir was also well tolerated in this population.
Our principal aim in estimating warfarin pharmacodynamics over the 5-day oseltamivir treatment course was to detect whether there was any change in blood coagulation. The advantage of calculating the net AUEC for INR and factor VIIa was that it enabled us to look for trends in either direction, i.e. towards an increase or a decrease in blood coagulation over the period. However, because the reports of increased bleeding in warfarinized patients from Canada and Europe had suggested reduced coagulation, we looked specifically at the maximum increase from baseline in INR and the maximum fall from baseline in factor VIIa. The absence of a pharmacodynamic interaction in our study is consistent with the lack of experimental evidence to date that oseltamivir interferes in any way with the normal coagulation process or with drugs that modify coagulation. Oseltamivir would not be expected to interfere with the normal metabolism of warfarin, as the two agents are not metabolized by the same pathway.
One possible explanation for the Canadian and European observations of altered response to warfarin in patients treated with oseltamivir for influenza is an effect of the virus itself. A case-control study in 90 warfarin-treated patients who received influenza vaccination found that INR values increased from 2.64 to 3.85 in about half the patients, with two having bleeding episodes [6]. However, in more than a dozen other published reports, mostly non-controlled studies in small numbers of vaccine recipients, significant changes in prothrombin time or INR value were not recorded [7–10]. There are some brief case reports of patients receiving warfarin that suggest a possible interaction with influenza vaccine, including a patient who experienced a serious, almost fatal, bleeding episode after receiving a ‘flu shot’[11]. The same group also reported the case of an elderly man who developed bleeding (haematemesis and melaena) within 10 days of receiving an influenza vaccination, and had a prothrombin time of 36 s [12]. Another case study involved a patient with raised INR values in 2 successive years during which influenza vaccination was administered [13], while a further case of a fatal intracranial bleed in a 64-year-old man was also recently reported [14]. In this latter report, the patient had an INR of 2.0 at the time of influenza vaccination and a history of stable INR values (range 1.4–4.7) during the previous 6 months. However, an INR of >15 was reported at the time of hospitalization, 4.5 weeks after vaccination. Therefore, there appears to be a possible association between warfarin and influenza vaccines, but no reports suggesting that any altered response to warfarin is an effect likely related to the use of oseltamivir.
Although the mean change from baseline in vitamin K1 concentrations during our study showed no difference between the treatment groups, it was notable that concentrations rose in the majority of the patients, substantially in some cases, and independent of treatment sequence. This might have been due to a change in the patients' diet during the study, although efforts were made to minimize intake of vitamin K1 and changes in diet during the study. It could also have been influenced by another factor associated with the patients' stay in the clinical trial unit, e.g. reduced physical activity. The mixed model analysis showed consequently that treatment period had a significant effect on the change in vitamin K1 from baseline.
In agreement with the lack of pharmacodynamic interaction between warfarin and oseltamivir in the current study, changes in the pharmacokinetics of either warfarin enantiomer ([R] or [S]) following oseltamivir treatment were not clinically relevant. Although there was no control group in the present study, we also assessed whether the interaction occurred in reverse by comparing pharmacokinetic parameters of oseltamivir and its active metabolite in this study with previously published and unpublished data [15–18]. The limited numbers of patients available and the historical nature of the comparison imply that these data should be interpreted with caution, but preliminary visual inspection indicated no identifiably meaningful effect of warfarin on the pharmacokinetics of oseltamivir.
Our study does not provide any evidence for a clinically relevant pharmacodynamic or pharmacokinetic interaction between oseltamivir and warfarin that could alter the anticoagulant effect of warfarin in individuals who take this drug at stable dosages for chronic vascular or cardiac conditions.
Acknowledgments
Medical writing support for the development of this article was provided by Roger Nutter and Stephen Purver at Gardiner-Caldwell Communications, Macclesfield, UK. Funding for this support was provided by F. Hoffmann-La Roche Ltd.
Competing interests
BD is an employee of Hoffmann-La Roche Inc. and holds non-voting stock in Hoffmann-La Roche Inc. PAB, SL-C and MB are employees of Roche Products Ltd. MB owns non-voting shares.
REFERENCES
- 1.Toovey S, Rayner C, Prinssen E, Chu T, Donner B, Thakrar B, Dutkowski R, Hoffmann G, Breidenbach A, Lindemann L, Carey E, Boak L, Gieschke R, Sacks S, Solsky J, Small I, Reddy D. Assessment of neuropsychiatric adverse events in influenza patients treated with oseltamivir: a comprehensive review. Drug Saf. 2008;31:1097–114. doi: 10.2165/0002018-200831120-00006. [DOI] [PubMed] [Google Scholar]
- 2.Roche. Tamiflu: summary of main product characteristics. electronic Medicines Compendium (eMC). 17-2-2009.
- 3.Dutkowski R, Thakrar B, Froehlich E, Suter P, Oo C, Ward P. Safety and pharmacology of oseltamivir in clinical use. Drug Saf. 2003;26:787–801. doi: 10.2165/00002018-200326110-00004. [DOI] [PubMed] [Google Scholar]
- 4.Pepper S, Murty H. Oseltamivir (Tamiflu) and warfarin: suspected increase in INR. Canadian Adverse Reaction Newsletter. 2006;16:1–2. [Google Scholar]
- 5.He G, Massarella J, Ward P. Clinical pharmacokinetics of the prodrug oseltamivir and its active metabolite Ro 64-0802. Clin Pharmacokinet. 1999;37:471–84. doi: 10.2165/00003088-199937060-00003. [DOI] [PubMed] [Google Scholar]
- 6.Paliani U, Filippucci E, Gresele P. Significant potentiation of anticoagulation by flu-vaccine during the season 2001–2002. Haematologica. 2003;88:599–600. [PubMed] [Google Scholar]
- 7.Lipsky BA, Pecoraro RE, Roben NJ, de BP, Delaney CJ. Influenza vaccination and warfarin anticoagulation. Ann Intern Med. 1984;100:835–7. doi: 10.7326/0003-4819-100-6-835. [DOI] [PubMed] [Google Scholar]
- 8.Gomolin IH, Chapron DJ, Luhan PA. Lack of effect of influenza vaccine on theophylline levels and warfarin anticoagulation in the elderly. J Am Geriatr Soc. 1985;33:269–72. doi: 10.1111/j.1532-5415.1985.tb07115.x. [DOI] [PubMed] [Google Scholar]
- 9.Bussey HI, Saklad JJ. Effect of influenza vaccine on chronic warfarin therapy. Drug Intell Clin Pharm. 1988;22:198–201. doi: 10.1177/106002808802200303. [DOI] [PubMed] [Google Scholar]
- 10.Raj G, Kumar R, McKinney WP. Safety of intramuscular influenza immunization among patients receiving long-term warfarin anticoagulation therapy. Arch Intern Med. 1995;155:1529–31. [PubMed] [Google Scholar]
- 11.Sumner HW, Holtzman JL, McClain CJ. Drug-induced liver disease. Geriatrics. 1981;36:83–96. [PubMed] [Google Scholar]
- 12.Kramer P, Tsuru M, Cook CE, McClain CJ, Holtzman JL. Effect of influenza vaccine on warfarin anticoagulation. Clin Pharmacol Ther. 1984;35:416–8. doi: 10.1038/clpt.1984.52. [DOI] [PubMed] [Google Scholar]
- 13.Beeley L, Cunningham H, Carmichael A, Brennan A. Bulletin for the West Midland Centre for Adverse Drug Reaction Reporting. 1991;33:19. [Google Scholar]
- 14.Carroll DN, Carroll DG. Fatal intracranial bleed potentially due to a warfarin and influenza vaccine interaction. Ann Pharmacother. 2009;43:754–60. doi: 10.1345/aph.1L413. [DOI] [PubMed] [Google Scholar]
- 15.Ward P, Massarella J, Dorr A, Ipe D. Clinical Study Report – Protocol NP15717. Study of the pharmacodynamics and pharmacokinetics of the neuraminidase inhibitor oseltamivir (GS4104) in the treatment of volunteers experimentally infected with human influenza B virus. Research Report No. W-144105, 1999 (Roche, data on file)
- 16.Davies B, Huson L. Clinical Study Report – Protocol WP20727. A single ascending dose and exploratory bioavailability study to investigate the safety, tolerability and pharmacokinetics of intravenous Ro 64-0796 in healthy volunteers. Research Report No. 1030591, 1997 (Roche, data on file)
- 17.Schentag JJ, Hill G, Chu T, Rayner C. Similarity in pharmacokinetics of oseltamivir and oseltamivir carboxylate in Japanese and Caucasian subjects. J Clin Pharmacol. 2007;47:689–96. doi: 10.1177/0091270007299761. [DOI] [PubMed] [Google Scholar]
- 18.Oo C, Barrett J, Dorr A, Liu B, Ward P. Lack of pharmacokinetic interaction between the oral anti-influenza prodrug oseltamivir and aspirin. Antimicrob Agents Chemother. 2002;46:1993–5. doi: 10.1128/AAC.46.6.1993-1995.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]