Comparison of the 5-Fluorouracil-Warfarin and Capecitabine-Warfarin Drug Interactions

Comparison of the 5-Fluorouracil–Warfarin andCapecitabine-Warfarin Drug Interactions

Sachin R. Shah, Pharm.D., Randall Martin, Pharm.D., Jonathan E. Dowell, M.D., andSarah M. Gressett Ussery, Pharm.D.

Study Objective. To compare the differences between the 5-fluorouracil–warfarinand capecitabine-warfarin drug interactions in patients receiving therapeuticdoses of warfarin.

Design. Retrospective medical record review.Setting. Large academic Veterans Affairs health care system.Patients. Twenty-four patients who received concomitant therapeutic doses of

warfarin and at least one dose of a fluoropyrimidine—5-fluorouracil orcapecitabine—between January 2004 and May 2008.

Measurements and Main Results. The primary outcome was mean change ininternational normalized ratio (INR) from baseline to end of 90-day studyperiod. Only patients who were taking warfarin before starting eitherfluoropyrimidine-based chemotherapy were included in the primaryanalysis. Of the 24 eligible patients, 15 (9 receiving 5-fluorouracil, 6receiving capecitabine) were taking warfarin before receiving eitherfluoropyrimidine. No significant differences were noted in the averageweekly warfarin dose or baseline INR between the 5-fluorouracil andcapecitabine groups. The mean change in INR for patients taking warfarinbefore fluoropyrimidine use was 4.62 in the 5-fluorouracil group comparedwith 5.11 in the capecitabine group (p=0.87). The capecitabine group hada similar proportion of patients achieving an INR above 9 while takingwarfarin compared with the 5-fluorouracil group (17% vs 22%). In thosepatients who required a warfarin dosage reduction, the dose was reduced by38% and 41% in the 5-fluorouracil and capecitabine groups, respectively.No significant differences in bleeding events were reported within 90 daysof concurrent use of either fluoropyrimidine with warfarin.

Conclusion. Our study suggests that in patients receiving concomitantwarfarin and a fluoropyrimidine, no significant differences in INR elevationor bleeding events were noted with 5-fluorouracil compared with capecitabine.

Key Words: capecitabine, drug interactions, fluoropyrimidine, 5-fluorouracil,5-FU, international normalized ratio, INR, warfarin, anticoagulation.(Pharmacotherapy 2010;30(12):1259–1265)

Concomitant administration of either 5-fluorouracil or capecitabine with therapeuticdoses of warfarin has been reported to causeelevation of the international normalized ratio(INR), and in a few cases, symptomatic bleeding.1–12

5-Fluorouracil and capecitabine belong to thefluoropyrimidine class of antineoplastic agents

and are often used interchangeably. Bothfluoropyrimidines exert their effect by inhibitingthymidylate synthetase, which results in athymine deficiency and ultimately cell death.13, 14

However, important pharmacokinetic andpharmacodynamic differences between the twoagents exist.

PHARMACOTHERAPY Volume 30, Number 12, 2010

5-Fluorouracil has been used for almost 50years to treat various cancers, including colon,head and neck, esophageal, gastric, and breastcancer.15 Due to variable and poor oralbioavailability, the drug is only administeredintravenously; it is metabolized in the liver bydihydropyrimidine dehydrogenase before beingexcreted renally. Capecitabine, which wasapproved in 1998 for use in advanced ormetastatic breast and colon cancers, is an oralprodrug of 5-fluorouracil that undergoesbiotransformation to 5-fluorouracil through athree-step enzymatic process.1, 13–15 The first twosteps occur in the liver, with the third stepoccurring in the tumor cell.13–15 This final step ismediated by the enzyme thymidine phosphorylase,which is expressed at a higher concentration inthe tumor cells.16

Approximately 11% of capecitabine users takewarfarin concomitantly, whereas the proportion5-fluorouracil users who take warfarin isunknown.12 Thromboembolic disease is commonin patients with cancer, and warfarin is frequentlyused to treat this complication. In addition,many patients with cancer may be receivingwarfarin for other comorbidities such as atrialfibrillation or genetic hypercoagulable states.Warfarin is extensively metabolized by thecytochrome P450 (CYP) enzyme system, specifi-cally by CYP2C9, 3A4, and 1A2 isoenzymes, aswell as others.1, 17, 18 Many drug interactionsaffecting the INR result from drug-inducedalterations of this CYP metabolism. A prospectivepharmacokinetic-pharmacodynamic study of theinteraction between warfarin and capecitabinesuggests involvement of this system, particularlyCYP2C9.18

The capecitabine package insert has a black-box warning regarding the drug interaction withwarfarin.19 The 5-fluorouracil package insert alsowarns of the drug interaction with warfarin.20

Both package inserts recommend close monitoringof patients who take either of these agents concur-

rently with warfarin. It is well established that 5-fluorouracil or capecitabine in combination withwarfarin can result in a significant increase in theINR value.1–12 However, data are lacking thatillustrate the magnitude of the rise in INR or thefrequency of the 5-fluorouracil–warfarin druginteraction. Most important, we found no head-to-head comparisons of the 5-fluorouracil–warfarin and capecitabine-warfarin drug inter-actions with regard to their frequencies andmagnitude.

5-Fluorouracil and capecitabine are commonlyused interchangeably in clinical practice,especially for the treatment of patients withcolorectal cancer. Therefore, it is important torecognize the differences in the drug-druginteraction characteristics between these twochemotherapy agents with warfarin. The purposeof this study was to evaluate the differencesbetween the 5-fluorouracil–warfarin andcapecitabine-warfarin drug interactions inpatients receiving therapeutic doses of warfarin.

Methods

This was a single-center, retrospective reviewof the medical records of all patients whoreceived warfarin and either 5-fluorouracil orcapecitabine concomitantly between January2004 and May 2008 at the Veterans Affairs NorthTexas Health Care System (Dallas, TX). Onlypatients aged 18 years or older who received atleast one dose of either 5-fluorouracil orcapecitabine and one dose of warfarin concomi-tantly were included in the study. Patients wereeligible for inclusion only if therapeutic doses ofwarfarin were used. Patients were excluded if noINRs were documented after the baseline INRvalue and if warfarin was being used for primaryprophylaxis or to maintain central catheterpatency. Eligible patients were then assigned toone of two groups based on which fluoro-pyrimidine—capecitabine or 5-fluorouracil—thepatient received. The study was approved by theinstitutional review board, and informed consentwas waived due to the retrospective nature of thestudy.

The primary outcome of this study was meanchange in INR from baseline to the end of a 90-day study period. Data were collected through acomputerized patient record system, using auniform data collection form. Demographicinformation collected included age, sex, height,weight, chemotherapy and warfarin indications,dosages and start dates of the fluoropyrimidines

1260

From the School of Pharmacy, Texas Tech UniversityHealth Sciences Center (Drs. Shah and Martin), VeteransAffairs North Texas Health Care System (Drs. Shah, Dowell,and Ussery), and University of Texas Southwestern MedicalCenter (Dr. Dowell), Dallas, Texas.

Presented as a poster at the American College of ClinicalPharmacy Spring Practice and Research Forum, Charlotte,North Carolina, April 23–27, 2010.

For reprints, visit http://www.atypon-link.com/PPI/loi/phco.For questions or comments, contact Sachin R. Shah,Pharm.D., BCOP, Texas Tech University Health SciencesCenter School of Pharmacy, 4500 South Lancaster Road,Building 7, Room 119-A, Dallas, TX 75216; e-mail:[email protected].

5-FU–WARFARIN VERSUS CAPECITABINE-WARFARIN DRUG INTERACTIONS Shah et al

and warfarin, presence of metastasis (in particu-lar, liver), target INR, and dates that warfarin andthe fluoropyrimidine were used concurrently.The following end point data were collected:date that an INR above 3 was reached, date thatan INR above 9 was reached, date and value ofhighest INR achieved, date and severity of anybleeding events, date and type of any thrombo-embolic events, date and cause of death ifapplicable, warfarin dosage reduction, warfarinchange to another anticoagulant, fluoropyrimidinechange to a different agent or discontinued, andrescue interventions (fresh frozen plasma, vitaminK) utilized.

Although this study included all patientsreceiving both warfarin and a fluoropyrimidineconcurrently, only patients who were takingwarfarin before fluoropyrimidine use wereincluded in the primary analysis.

Statistical Analysis

An a priori power analysis was conducted todetermine an appropriate sample size. Based on

a power of 80%, a two-tailed � value of 0.05, anda standard deviation of 1, we determined that atleast six patients in each fluoropyrimidine groupwould be required to detect a mean difference of2 in INR elevation between groups. All statisticalanalyses were performed using Minitab, release15 (Minitab Inc., State College, PA) statisticalsoftware. Statistical significance was defined as ap value less than 0.05 for all data analyses. Theprimary end point of mean difference in maximumINR elevation between groups was analyzed byusing the Student t test. Additional continuousvariables were also analyzed by using Student t orMann Whitney U test, and nominal variableswere analyzed by using the Pearson �2 or Fisherexact tests, where applicable.

Results

Thirty-nine patients were identified after theinitial screening. On further review, four patientswere excluded because they had not received anychemotherapy. Nine were excluded because theyhad not received warfarin and a fluoropyrimidine

1261

Table 1. Baseline Demographic and Clinical Characteristics of the Study Patients

Patients Taking Warfarin Before Patients Taking Warfarin Before or AfterConcomitant Fluoropyrimidine Use Concomitant Fluoropyrimidine Use5-Fluorouracil Capecitabine 5-Fluorouracil Capecitabine

Group Group Group GroupCharacteristic (n=9) (n=6) (n=15) (n=9)Age (yrs)

Mean 62 62 62 63Median 60 60 60 61

Body mass index, mean (kg/m2)a 27.6 28.5 26.1 28.3Weekly warfarin dose, mean (mg) 39 31 36 35Baseline INR

Mean ± SD 1.68 ± 0.76 2.02 ± 0.61 1.58 ± 0.72 1.76 ± 0.62Median 1.40 1.96 1.30 1.70

No. (%) of PatientsMale 8 (89) 5 (83) 14 (93) 8 (89)Chemotherapy indication

Colorectal cancer 8 (89) 5 (83) 14 (93) 8 (89)Head and neck cancer 1 (11) 0 (0) 1 (7) 0 (0)Breast cancer 0 (0) 1 (17) 0 (0) 1 (11)

Metastatic cancer 5 (56) 6 (100) 8 (53) 9 (100)Liver metastasis 2 (22) 3 (50) 4 (27) 5 (56)

Warfarin indicationAtrial fibrillation 3 (33) 2 (33) 4 (27) 3 (33)Deep vein thrombosis 5 (56) 1 (17) 7 (47) 3 (33)Pulmonary embolism 1 (11) 2 (33) 4 (27) 2 (22)Mitral valve replacement 0 (0) 1 (17) 0 (0) 1 (11)

Anticoagulation managementa

Anticoagulation clinic 5 (56) 3 (50) 11 (73) 4 (44)Primary care clinic 4 (44) 3 (50) 4 (27) 5 (56)

Concurrent antibiotic therapya 1 (11) 2 (33) 3 (20) 2 (22)aDifferences between groups were not statistically significant.


concurrently. An additional two patients wereexcluded: one patient did not have any INRsdocumented in the medical record, and the otherdid not have warfarin therapy intent and targetINR documented. Thus, 24 eligible patients wereidentified for the overall analysis. Of these, 15were taking warfarin before receiving fluoro-pyrimidine-based chemotherapy.

Warfarin therapy was monitored by either theanticoagulation clinic or primary care clinic. Nochanges were required in the warfarin adminis-tration schedule based on the chemotherapyadministration cycle.

Demographic characteristics of the 24 eligiblepatients (those taking warfarin before or afterconcomitant fluoropyrimidine use) and of justthe 15 patients taking warfarin before fluoro-pyrimidine use are shown in Table 1. These dataclarify the relevant impact of 5-fluorouracil orcapecitabine on patients already receivingwarfarin therapy. No significant differences atbaseline were noted between the 5-fluorouraciland capecitabine groups: median age was

approximately 60 years, primary indication forchemotherapy was colorectal cancer, and primaryindications for warfarin therapy were deep veinthrombosis and pulmonary embolism. Inaddition, no significant differences in the averageweekly warfarin dose and baseline INR werenoted between the 5-fluorouracil and capecitabinegroups. In the capecitabine group, all patientshad received the drug for 2 weeks, followed by 1week without the drug, every 21 days.

The mean change in INR from baseline to theend of a 90-day study period for patients takingwarfarin before fluoropyrimidine use was 4.62 forthe 5-fluorouracil group compared with 5.11 forthe capecitabine group (p=0.87). This findingwas consistent in all eligible patients (4.05 forthe 5-fluorouracil group vs 5.51 for thecapecitabine group, p=0.51), regardless of whenwarfarin was started, either before or afterfluoropyrimidine use (Table 2). The peak INRachieved was 19.1 in the 5-fluorouracil groupand 16.71 in the capecitabine group (Figure 1).For patients taking warfarin before fluoro-

1262

Table 2. International Normalized Ratio Changes and Interventions Required

Patients Taking Warfarin Before Patients Taking Warfarin Before or AfterConcomitant Fluoropyrimidine Use Concomitant Fluoropyrimidine Use5-Fluorouracil Capecitabine 5-Fluorouracil Capecitabine

Group Group Group GroupVariable (n=9) (n=6) (n=15) (n=9)INR change from baseline

Mean ± SD 4.62 ± 6.02 5.11 ± 5.23 4.05 ± 4.72 5.51 ± 5.39Median 2.90 4.15 2.80 3.90

Peak INRMean ± SD 6.30 ± 5.99 7.13 ± 5.13 5.63 ± 4.71 7.27 ± 5.29Range 2–19.10 2.67–16.71 2–19.10 2.3–16.71

Days to Peak INRMean 46 32 41 33Median 35 28 35 28

Days to INR > 3Mean 23 28 21 27Median 23 26 20 26

Weekly warfarin dose, 31 21 31 26mean (mg)a, b

No. (%) of PatientsINR > 3 5 (56) 5 (83) 10 (67) 7 (78)INR > 9 2 (22) 1 (17) 2 (13) 2 (22)Interventionc

Warfarin dose reduced 4 (44) 4 (67) 9 (60) 5 (56)Changed to alternative 1 (11) 1 (17) 1 (7) 2 (22)anticoagulant

Rescue therapy needed 2 (22) 1 (17) 3 (20) 2 (22)Chemotherapy changed 0 (0) 1 (17) 0 (0) 1 (11)

aAfter concurrent administration with 5-fluorouracil or capecitabine.bPatients were excluded from the analysis if warfarin was changed to an alternate anticoagulant. Differences between groups were notstatistically significant.cSome patients required more than one intervention.


pyrimidine use, the capecitabine group had ahigher proportion of patients achieving an INRabove 3 while taking warfarin (83% vs 56%,p=0.58). In addition, the median time to reachpeak INR was shorter in the capecitabine group:28 days compared with 35 days for 5-fluorouracil(p=0.32). Within 90 days of concurrentlyreceiving chemotherapy with warfarin, bleedingwas reported in 3 (33%) of 9 patients in the 5-fluorouracil group and in 1 (17%) of 6 patients inthe capecitabine group (p=0.58). One of thebleeding events in the 5-fluorouracil group wasconsidered to be major.

Deep vein thrombosis was reported in onepatient while concurrently receiving warfarin and5-fluorouracil. None of the patients in eithergroup died within their 90-day study period ofconcurrently receiving warfarin and a fluoro-pyrimidine. The significant interventions thatwere required due to the warfarin-fluoro-pyrimidine interactions are shown in Table 2. Inthe primary analysis of patients taking warfarinbefore fluoropyrimidine use, warfarin dosagereduction was required in four patients (44%) inthe 5-fluorouracil group compared with fourpatients (67%) in the capecitabine group. Inthose patients who required a warfarin dosagereduction, the dose of warfarin was reduced by38% and 41% in the 5-fluorouracil andcapecitabine groups, respectively.

For all patients, regardless of when warfarinwas started, three had their warfarin changed toan alternate anticoagulant. Two of these patientswere taking warfarin therapy before starting

chemotherapy. In both of these patients, warfarintherapy was changed to an alternate anticoagulantdue to an INR value of 19.1 (5-fluorouracilgroup) and 15.2 (capecitabine group). In thethird patient (capecitabine group), warfarin waschanged to low-molecular-weight heparin inanticipation of the drug interaction. For allpatients, rescue therapy was needed in fivepatients, and chemotherapy was changed in onepatient in the capecitabine group.

Discussion

This study systematically compared changes inINR values due to the drug interaction of 5-fluorouracil or its prodrug capecitabine withwarfarin. To our knowledge, this is the firstretrospective evaluation of the effects of 5-fluorouracil on warfarin dosed to target atherapeutic INR. The study confirms thesignificant risk of the drug-drug interaction whenwarfarin is taken concurrently with 5-fluorouracil or capecitabine. Patients receiving afluoropyrimidine were at 67% risk of having anINR above 3 and at 20% risk of having an INRabove 9. The study suggests a trend toward ahigher mean INR change and a shorter time topeak INR with capecitabine, regardless of whenpatients began warfarin (before or afterfluoropyrimidine use). The INR increased 2.8times from baseline in the 5-fluorouracil groupcompared with 2.5 times from baseline in thecapecitabine group. These trends, however, arenot statistically significant and may not represent

1263

Figure 1. Changes in international normalized ratio (INR) from baseline in patients taking warfarin before concomitant 5-fluorouracil (nine patients) or capecitabine (six patients) use.

20

15

10

5

0

3

Baseline INR Peak INR Baseline INR Peak INR

5-Fluorouracil Capecitabine

INR

Valu

e


a clinically significant difference between the 5-fluorouracil and capecitabine groups. Inaddition, the weekly dose reduction required ofwarfarin after concurrent fluoropyrimidine usewas similar between groups. This study suggeststhat capecitabine, before its conversion to 5-fluorouracil, may not have a significant inter-action with warfarin, since the mean change inINR was not statistically significantly differentbetween the 5-fluorouracil and capecitabinegroups.

As expected, the time to peak INR was shorterfor patients receiving warfarin at the therapeutictarget dose than that reported in studies ofminidose (prophylaxis dose) warfarin with 5-fluorouracil.21–23 To our knowledge, a study of 5-fluorouracil in combination with warfarin dosedwith therapeutic intent has not been published.

The effects of patients receiving warfarin incombination with capecitabine have beenevaluated retrospectively.1 Twelve of 21 patientswere taking warfarin before initiating capecitabine;however, the study failed to break down theresults by timing of warfarin initiation (before vsafter capecitabine). The 21 patients received anaverage weekly warfarin dose of 18.8 mg duringcapecitabine therapy. Maximum INRs rangedfrom 1–11.5, and 29% of patients requiredwarfarin dose reductions. The likely reasons forthe lower INR range and lower percentage ofwarfarin dose reductions (compared with ourstudy) were lower weekly doses of warfarin andthe inclusion of patients receiving warfarin atprophylactic doses. Ten of the 21 patients in thatreport were receiving warfarin to maintain portpatency. The study did not report the percentdose reductions in patients who had theirwarfarin dose reduced. Similar to our study,metastatic liver involvement did not have asignificant impact on the mean INR change of thepatients (data not reported).

Conclusion

This study supports the findings of previouscase reports and retrospective evaluations ofwarfarin-fluoropyrimidine drug-drug interactions.In patients receiving concomitant warfarin and afluoropyrimidine, we found no significant differ-ences in INR elevation from baseline through 90days, time to peak INR, or bleeding eventsbetween the 5-fluorouracil and capecitabinegroups. A warfarin dose reduction of approxi-mately 40% is required after initiation of 5-fluorouracil or capecitabine in patients at risk for

the interaction. Clinicians should avoidchanging patients’ 5-fluorouracil therapy tocapecitabine, or vice versa, if the primaryconcern is a drug-drug interaction with warfarin.The high occurrence of supratherapeutic INRs inour study suggests that low-molecular-weightheparin may be a better option than warfarin, iffeasible, in patients scheduled to receive 5-fluorouracil or capecitabine. The specificmechanism of the warfarin-fluoropyrimidinedrug-drug interaction is poorly understood andshould be further investigated.

Acknowledgment

We would like to thank Melissa Eder for her editorialassistance in the preparation of this manuscript.

References1. Shah HR, Ledbetter L, Diasio R, Saif MW. A retrospective

study of coagulation abnormalities in patients receivingconcomitant capecitabine and warfarin. Clin Colorectal Cancer2006;5:354–8.

2. Davis DA, Fugate SE. Increasing warfarin dosage reductionsassociated with concurrent warfarin and repeated cycles of 5-fluorouracil therapy. Pharmacotherapy 2005;25:442–7.

3. Brown MC. An adverse interaction between warfarin and 5-fluorouracil. A case report and review of the literature.Chemotherapy 1999;45:392–5.

4. Brown MC. Multisite mucous membrane bleeding due to apossible interaction between warfarin and 5-fluorouracil.Pharmacotherapy 1997;17:631–3.

5. Carabino J, Wang F. International normalized ratio fluctuationwith warfarin-fluorouracil therapy [letter]. Am J Health SystPharm 2002;59:875.

6. Kolesar JM, Johnson CL, Freeberg BL, Berlin JD, Schiller JH.Warfarin-5-FU interaction—a consecutive case series.Pharmacotherapy 1999;19:1445–9.

7. Yildirim Y, Ozyilkan O, Akcali Z, Basturk B. Drug interactionbetween capecitabine and warfarin: a case report and review ofthe literature. Int J Clin Pharmacol Ther 2006;44:80–2.

8. Isaacs K, Haim N. Adverse interaction between capecitabineand warfarin resulting in altered coagulation parameters andbleeding: case report and review of the literature. J Chemother2005;17:339–42.

9. Janney LM, Waterbury NV. Capecitabine-warfarin interaction.Ann Pharmacother 2005;39:1546–51.

10. Copur MS, Ledakis P, Bolton M, et al. An adverse interactionbetween warfarin and capecitabine: a case report and review ofthe literature. Clin Colorectal Cancer 2001;1:182–4.

11. Buyck HC, Buckley N, Leslie MD, Plowman PN. Capecitabine-induced potentiation of warfarin [letter]. Clin Oncol (R CollRadiol) 2003;15:297.

12. Yood MU, Quesenberry CP Jr, Alford SH, et al . Anobservational study examining the impact of capecitabine onwarfarin antithrombotic activity and bleeding complications.Curr Med Res Opin 2006;22:307–14.

13. Blum JL. The role of capecitabine, an oral, enzymaticallyactivated fluoropyrimidine, in the treatment of metastatic breastcancer. Oncologist 2001;6:56–64.

14. Rustum YM. Thymidylate synthase: a critical target in cancertherapy? Front Biosci 2004;9:2467–73.

15. U.S. Food and Drug Administration. Drugs@FDA: FDAapproved drug products. Available from www.fda.gov/search/databases.html. Accessed September 12, 2009.

16. Miwa M, Ura M, Nishida M, et al. Design of a novel oralfluoropyrimidine carbamate, capecitabine, which generates 5-

1264


fluorouracil selectively in tumours by enzymes concentrated inhuman liver and cancer tissue. Eur J Cancer 1998;34:1274–81.

17. Giunta G. Warfarin–5-FU interactions [letter]. Ann Oncol2006;17:176.

18. Camidge R, Reigner B, Cassidy J, et al. Significant effect ofcapecitabine on the pharmacokinetics and pharmacodynamicsof warfarin in patients with cancer. J Clin Oncol 2005;23:4719–25.

19. Genentech. Xeloda (capecitabine) package insert. South SanFrancisco, CA; 2009.

20. APP Pharmaceuticals. Fluorouracil injection package insert.Schaumburg, IL; 2008.

21. Magagnoli M, Masci G, Castagna L, Morenghi E, Santoro A.High incidence of INR alteration in gastrointestinal cancerpatients treated with mini-dose warfarin and 5-fluorouracil–based regimens. Ann Oncol 2006;17:174–6.

22. Magagnoli M, Masci G, Castagna L, et al. High incidence ofhaemostatic interference in cancer patients treated withFOLFOX regimen and concomitant minidose of warfarin. Br JHaematol 2005;129:709–10.

23. Magagnoli M, Castagna L, Masci G, Santoro A. Low-dosewarfarin prophylaxis for catheter-associated thrombosis incancer patients. Can it be safely associated with 5-fluorouracil–based chemotherapy? Oncologist 2004;9:594–5.

1265

Documents

Comparison of the 5-Fluorouracil-Warfarin and Capecitabine-Warfarin Drug Interactions