A Comprehensive Review of Potential Warfarin-Fruit Interactions
A Comprehensive Review of Potential Warfarin-Fruit Interactions
Abstract
Purpose: The aim of this review is to discuss possible interactions that may occur between warfarin and fruit products. Methods: A
literature search was conducted using the search terms: ‘‘warfarin (Coumadin1) and fruit interactions, warfarin and fruit, warfarin
and fruit juice, case reports and clinical trials’’. Results: A total of 23 citations (15 case reports and 7 controlled clinical trials) were
reviewed. The majority of cases involved cranberry products, while pomegranate juice, avocado, grapefruit juice, mango, and
papain were also implicated in reports of suspected warfarin-fruit interactions. Cranberry juice was also the most frequently
studied fruit product. Other fruit products evaluated with warfarin in controlled clinical trials were cranberry concentrate and
grapefruit juice. Conclusion: Although a number of case reports have been published that suggest warfarin has the potential to
interact with several fruit products, it is difficult to determine their relevance, as scientific evidence is scarce. Until further
information is available, clinicians may want to encourage patients to consume cranberry products and grapefruit juice in small to
moderate quantities and to inquire about the recent consumption of mangos, pomegranate juice, and avocados when taking a
dietary history or when assessing possible causes for international normalized ratio (INR) instability.
Keywords
ambulatory care, medication safety
Unknown or
Questions Yes No NA
One fruit product, cranberry juice, has been frequently pur- between warfarin and fruit products in humans. In addition, the
ported to interact with warfarin. In October 2004, the United fruit products had to be currently recognized by the United
Kingdom’s Medicines and Healthcare Products Regulatory States Department of Agriculture. Non-English language liter-
Agency had received several reports of a suspected cranberry ature was excluded, and references from published articles
juice–warfarin interaction and issued a statement that patients were reviewed for additional sources.
taking warfarin should avoid cranberry products unless the Currently, there is not a standardized assessment tool for
benefits clearly outweigh the risks.10 One year later, the pre- determining the probability of a drug–food interaction. There-
scribing information for Coumadin1 (warfarin sodium tablets, fore, to establish the likelihood of a warfarin-fruit interaction in
USP) was changed in the United States to include a statement the case reports reviewed, the Drug Interaction Probability
that patients taking warfarin should avoid cranberry juice and Scale (DIPS) was utilized.13 The DIPS is a modified version
cranberry products.11 These warnings were later removed in of the Naranjo adverse drug reaction probability scale14 and
2012, and currently there are no recommendations regarding is intended to evaluate the probability of a drug–drug interac-
the use of cranberry products in the prescribing information for tion occurring in a specific patient. For the purpose of this lit-
Coumadin. Adding to the uncertainty are popular drug informa- erature review, the DIPS was modified so that warfarin was
tion resources, such as Micromedex1, that lists major drug– labeled as the object drug, and the fruit products purported to
food interactions between warfarin and some fruit products interact with warfarin were considered the precipitant drug.
(eg, cranberry juice and pomegranate juice), with little to no An example of the modified DIPS form used in this review is
scientific evidence to support the claims.12 A clearer under- located in Table 1. In each report, the probability of an interac-
standing of the literature is necessary to help clinicians assess tion was rated as highly probable (total score > 8), probable
the potential for a warfarin-fruit interaction and advise patients (total score ¼ 5-8), possible (total score ¼ 2-4), or highly
accordingly. In this article, possible interactions between war- doubtful (total score < 2).
farin and various fruit products commonly consumed in the
United States are reviewed.
Results
Of the 10 307 citations identified, 23 (15 case reports and 7
Methods controlled clinical trials) met the criteria and were eligible for
To further examine which fruit products may interact with war- review. The majority of cases involved cranberry products (8
farin, a literature search was conducted within the date range of cases), while pomegranate juice (2 cases), avocado fruit (2
January 1950 to January 2014, using Medline (PubMed), cases), grapefruit juice (1 case), mango fruit (1 case), and
OvidSP, IDIS/Web, International Pharmaceutical Abstracts, papain (1case) were also implicated in reports of suspected
and The Cochrane Library. The following MeSH terms and key warfarin-fruit interactions. Thus far, cranberry juice, cran-
words were used: ‘‘warfarin (Coumadin) and fruit interactions, berry concentrate, and grapefruit juice are the only fruit prod-
warfarin and fruit, warfarin and fruit juice, case reports and ucts to be evaluated with warfarin in controlled clinical trials.
clinical trials’’. Articles were considered eligible for evaluation Of the trials, 5 involved cranberry juice, 1 used cranberry
if they contained original information regarding interactions concentrate, and 1 study examined the effects of grapefruit
Norwood et al 563
Abbreviations: CJ, cranberry juice, CC, cranberry concentrate capsules, CS, cranberry sauce; CYP, cytochrome P450; GJ, grapefruit juice; INR, international
normalized ratio; No., number; oz, ounces; pts, patients; w/, with.
juice on warfarin. Although cranberry concentrate capsules weeks after he started drinking cranberry juice, he was admitted
were shown to affect the pharmacodynamic properties of to the hospital with an INR >50 and later died of gastrointestinal
warfarin, neither cranberry juice nor grapefruit juice demon- and pericardial hemorrhage. Prior to the admission, the patient’s
strated a statistically significant interaction with warfarin in INR had been stable and he was not taking any medications other
any of the other trials. than digoxin, phenytoin, warfarin, and cephalexin for a chest
The case reports and clinical trials reviewed are summar- infection. Griffiths et al later published a more detailed account
ized subsequently. Patient considerations are also provided to of this case several years later and noted that prior to his death
help aid pharmacists and other health care professionals in the patient had been drinking 2 cups (300-400 mL) of pure cran-
counseling patients who are on warfarin. In addition, a brief berry juice daily for approximately 6 weeks.16 It was also men-
overview of suspected warfarin-fruit interactions is listed in tioned that the patient was compliant with his medications and
Table 2. had been using a tablet dispenser. According to the autopsy
report, the patient had a fatal hemorrhage into the pericardial sac
Discussion and gut. The authors of this report mentioned that hemorrhaging
in 2 different sites was indicative of a systemic coagulation
Cranberry (Vaccinium macrocarpon) abnormality most likely due to an increase in the anticoagulant
As mentioned, cranberry products have been the most fre- effect of warfarin. The autopsy report also showed that hemosi-
quently reported fruit products to interact with warfarin. Eight derin (iron pigment) and an organizing fibrin clot were present
different case reports describing an interaction between war- on the outer surface of the heart which suggests that the bleeding
farin and cranberry products have been published.15-23 in the pericardial sac was not spontaneous and must have
Although an exact mechanism to explain this potential drug– occurred over several days before the patient died. This finding
food interaction has yet to be determined, it has been sug- was consistent with the theory that the anticoagulant effect of
gested that cranberries may increase the anticoagulant effect warfarin was potentiated in this patient. In addition, information
of warfarin by inhibiting CYP3A4 and CYP2C9, 2 of the pri- from the toxicology report revealed that although the patient had
mary enzymes responsible for warfarin’s metabolism.24-26 In a massive hemorrhage and received 5 L of intravenous (IV)
addition, cranberries contain salicylic acid, which may fluids, warfarin still remained at detectable levels in his system.
potentiate the anticoagulant effect of warfarin by inhibiting In the absence of an overdose, this could indicate that the meta-
platelet aggregation and by displacing warfarin from protein bolism of warfarin was impaired. Based on the DIPS, it is pos-
binding sites.27,28 sible that a warfarin–cranberry juice interaction occurred in
this case (DIPS score ¼ 3 [þ1 þ 0 þ 0 þ 1 þ 0 þ 0 1 þ
Case reports. In the first case, Suvarna et al described an elderly 1 þ 1 þ 0]).
male in his 70s on warfarin who developed a poor appetite and In the next report, a 69-year-old male taking warfarin was
started consuming almost nothing but cranberry juice.15 Six admitted to the hospital in preparation for a sigmoid colectomy
564 Journal of Pharmacy Practice 28(6)
and bladder repair and was found to have an INR of 12.17 After restart warfarin therapy at a lower dose. After 2 weeks, the
holding multiple doses of warfarin and receiving vitamin K, his patient’s warfarin dose was titrated back to his regular weekly
INR eventually stabilized at 2 and he had surgery without any dose and his INR stabilized 1 week later at 2.4. Although the
complications. Several days after his surgery and after restart- DIPS indicates there was a possible warfarin–cranberry juice
ing warfarin, his INR once again increased to 8 and then to 11 interaction in this case, the likelihood of an interaction was
and was accompanied by frank hematuria from his catheter and weakened by the patient taking an extra dose of warfarin (DIPS
postrectal bleeding from the anastomosis site. Besides war- score ¼ 3 [þ1 þ 0 þ 0 þ 1 þ 1 þ 0 1 þ 0 þ 1 þ 0]).
farin, the only other medication the patient was taking was In this case, a 75-year-old caucasian man who had been on
digoxin and acetaminophen/codeine (postsurgery). The patient warfarin for 10 months went to a follow-up visit at an anticoa-
was not known to have any bleeding complications in the past gulation clinic 1 week after thanksgiving and was noted to have
and denied any changes to his medication regimen or diet but an INR of 4.8.20 After extensive questioning, the patient admit-
did admit to drinking almost 2 L/d of cranberry juice over the ted to eating *113 g of cranberry sauce daily for 1 week prior
2 weeks prior to his hospitalization to prevent urinary tract to the visit. He denied any alcohol consumption or changes to
infections. He was instructed to discontinue drinking cranberry his diet, health, and medication regimen (calcium carbonate,
juice, and 3 days later, his INR decreased to 3 and all bleeding cyanocobalamin, digoxin, cholecalciferol, metoprolol tartrate,
episodes stopped. According to the DIPS, it is probable that an simvastatin, and furosemide). The patient was instructed to dis-
interaction between warfarin and cranberry juice led to the continue eating cranberry sauce, hold warfarin for 2 days, and
events in this case (DIPS score ¼ 6 [þ1 þ 0 þ 0 þ 1 þ 1 þ then to decrease his warfarin dose from 22.5 mg/week to 20
2 1 þ 0 þ 1 þ 1]). mg/week. Approximately 7 days later, the patient’s INR was
The third case describes a 71-year-old man on warfarin within therapeutic range at 2.2 and continued to be stable over
therapy for stroke prophylaxis who was admitted to the hos- the next month. Based on the DIPS, it is possible that an inter-
pital with a 2 day history of hemoptysis, hematochezia, and action between warfarin and cranberry sauce resulted in the
shortness of breath.18 Upon admission, his INR and PT were patient’s supratherapeutic INR (DIPS score ¼ 3 [þ1 þ 0 þ 0
found to be >18 and >120 seconds, respectively. The patient þ 1 þ 0 þ 0 1 þ 0 þ 1 þ 1]).
denied any alcohol use or changes to his diet and health but In a recently documented case, a 64-year-old male who was
did state that 2 weeks prior to his admission, he began drink- stabilized on warfarin therapy (60 mg/week for > 1 year) was
ing 24 ounces (oz; 710 mL) of cranberry juice daily as a found to have a supratherapeutic INR of 5.5 during a clinic
source of vitamin C. While hospitalized, the patient received visit.21 According to the patient, he was only taking his usual
packed red blood cells, fresh frozen plasma, subcutaneous medications which included captopril, furosemide, gabapentin,
vitamin K, and gatifloxacin for a presumed chronic bronchitis pravastatin, metoprolol, sertraline, and hydrocodone 5 mg/
exacerbation. After several days, his INR eventually acetaminophen 500 mg (2 tablets at bedtime). He also denied
decreased to 2.6 and he was restarted on warfarin upon dis- any alcohol or tobacco use and had been eating a consistent
charge. After alleviating cranberry juice from his diet, he had amount of vitamin K-containing foods (a serving of coleslaw
2 subsequent INRs between 2 and 3. Although a cranberry or an iceberg salad once per week) prior to the elevated INR.
juice–warfarin interaction appears to be a likely contributor Apparently, the only change in his dietary habits was that he
to the elevated INR in this patient, another potential factor started drinking 2 glasses of cranberry juice daily over the past
that must be considered was the presence of an active infec- month. The patient was instructed to stop drinking cranberry
tion. Some infections can reduce the activity of several drug- juice, hold 2 doses of warfarin, and then to resume his normal
metabolizing enzymes, including CYP2C9, which could dose of 60 mg per week. Two weeks later, his INR was elevated
potentially enhance the anticoagulant effect of warfarin.29 at 3.6 which was attributed to a recent increase in his hydroco-
Using the DIPS, a cranberry juice–warfarin interaction was done/acetaminophen dose from 2 tablets/d (1000 mg of aceta-
determined to be possible in this report (DIPS score ¼ 3 minophen) to 6 tablets/d (3000 mg of acetaminophen). Use of
[þ1 þ 0 þ 0 þ 1 þ 1 þ 0 1 þ 0 þ 1 þ 0]). the DIPS indicated a possible warfarin–cranberry juice interac-
Another case report described a 78-year-old male on chronic tion for the initial elevated INR in this case (DIPS score ¼ 4
warfarin therapy for atrial fibrillation that was found to have an [þ1 þ 0 þ 0 þ 1 þ 1 þ 0 1 þ 0 þ 1 þ 1]).
INR of 6.45 during a regular follow-up visit with his provi- In the next report, a 46-year-old woman with an average INR
der.19 The patient denied alcohol use or any changes to his of 2.0 (INR range was 1.6-2.2 for 4 months) experienced an ele-
medication regimen (warfarin, isosorbide mononitrate, vated INR of 4.6 after drinking approximately 1.5 quarts (1420
digoxin, flunisolide, and loratadine) but did apparently take mL) of cranberry juice cocktail daily for 2 days.22 Although the
an extra dose of warfarin by accident during the previous week. patient reported using acetaminophen rarely, she denied any
The patient also admitted to taking a homemade cough prepara- alcohol or herbal medication use or any changes in her medica-
tion consisting of honey, vinegar, and water. The only known tion regimen. She was instructed to hold warfarin for 1 day, dis-
change in his diet was that he began drinking nearly a half- continue drinking cranberry juice, and then to resume her
gallon (1893 mL) of cranberry–apple juice during the previous normal weekly dose of warfarin. Two weeks later her INR was
week. He was instructed to hold multiple doses of warfarin, dis- 2.3 and continued to remain within the range of 1.4 to 2.5 (aver-
continue drinking the cranberry juice cocktail, and then to age INR 2.1) for the next 3 months. At a follow-up visit, the
Norwood et al 565
patient’s INR had once again increased, this time to 6.5. She therapy for at least 3 months, randomly consumed either
admitted to drinking approximately 2 quarts (1893 mL) of cran- 250 mL of cranberry juice or placebo daily for 1 week fol-
berry juice cocktail for 3 to 4 days prior to the measurement. lowed by a 1-week washout period. After the washout period,
After holding warfarin for 3 days and discontinuing cranberry subjects then drank the other beverage for 1 week so that all
juice, her INR decreased to 1.86. She then resumed her usual participants consumed either placebo or cranberry juice for a
warfarin dose and 1 week later her INR was 3.2. No signs or total of 1 week each. During the study, patients were
symptoms of bleeding were noticed at any time during the instructed to continue their normal warfarin regimen and not
supratherapeutic INR results. The possibility of a drug–food to change any of their dietary or exercise habits. INR values
interaction was strengthened in this case due to an elevation in did not differ between the 2 groups at baseline (2.28 þ 0.53
the patient’s INR on 2 separate occasions after consuming cran- for the cranberry juice group and 2.16 þ 0.40 for the placebo
berry juice. Based on the DIPS, a probable interaction between group), on day 7 (2.23 þ 0.53 for the cranberry juice group
warfarin and cranberry juice occurred in this report (DIPS and 2.16 þ 0.40 for the placebo group), or at any point dur-
score ¼ 6 [þ1 þ 0 þ 0 þ 1 þ 1 þ 2 1 þ 0 þ 1 þ 1]). ing the study. It was thereby concluded that the consumption
The most recent case to date involved an 85-year-old of 250 mL daily of cranberry juice does not adversely affect
woman on chronic warfarin therapy who’s INR was found to INR levels in patients taking warfarin.
be 5.1, 4 days after ingesting approximately 2 tablespoons of Next, in a randomized, crossover study, the effects of cran-
homemade cranberry sauce.23 Two months prior, her INR was berry juice on the pharmacodynamics of warfarin and the phar-
2.3 (target range 2-3) and her total weekly warfarin dose was macokinetics of racemic R-S warfarin, tizanidine, and
increased from 50 mg to 52.5 mg. The patient’s chronic med- midazolam were examined.31 For 10 days, 10 healthy volun-
ications included lisinopril, hydrochlorothiazide, metoprolol teers with a baseline INR of <1.3 ingested 200 mL of cranberry
succinate extended release, digoxin, and 2 homeopathic reme- juice (from concentrate) or water 3 times daily, except on day
dies (Naja Tripudia and Cactus Grandiflorus). She denied any 5, when cranberry juice or water was ingested 4 times through-
changes in her medications or dietary habits and did not use out the day. Additionally, on day 5, a cocktail of 10 mg of race-
tobacco or alcohol. She was instructed to hold warfarin for 4 mic warfarin, 1 mg of tizanidine, and 0.5 mg of midazolam was
days and then to decrease her total warfarin dose to 50 mg ingested simultaneously with the 200 mL of water or cranberry
weekly. Eleven days later, her INR was 1.7 and she was juice. Results revealed that cranberry juice reduced the area
instructed to continue on the same dose of warfarin. Seventeen under the plasma concentration–time curve (AUC0-1) of
days later, her INR was found to be 6.7 (using a point of care S-warfarin by 7% (P ¼ .051) and slightly shortened its elimina-
finger testing device) and 7.1 (confirmed by a laboratory per- tion half-life (t½) from 40.3 hours (control phase) to 36.4 hours
formed venipunture). According to the patient, she did not (P < .05). It did not significantly increase the peak plasma con-
make any changes to her warfarin dose but did again consume centration (Cmax), time to Cmax (tmax), AUC0-1, or t½ of R-S
cranberry sauce 3 days earlier. She was instructed to hold warfarin, tizanidine, and midazolam. In addition, the INR did
3 doses of warfarin and then to continue taking the same dose not differ during the cranberry juice phase compared to the con-
(50 mg/weekly). Her INR was 1.3, 5 days later and her weekly trol phase. Based on these results, the authors determined that
warfarin dose was slightly decreased to 47.5 mg. For the next cranberry juice does not significantly inhibit the in vivo activ-
2 months, her INR results were therapeutic at 2.1 and 2.7, ities of CYP2C9, CYP1A2, or CYP3A4 and does not alter the
respectively, using a weekly warfarin dose of 45 mg. Over the pharmacodynamic effect of warfarin.
next year, her INR levels ranged from 2 to 3 at a warfarin dose In another study, the impact of garlic and cranberry on the
of 47.5 mg weekly. It is unknown if the 2 homeopathic medi- pharmacokinetics and pharmacodynamics of warfarin was
cines used by the patient in this case may have interacted with assessed in an open-label, randomized, crossover clinical
warfarin. However, because she was reportedly taking them on trial.32 Twelve healthy males between the ages of 20 and
a daily basis and most of her INR levels during that time were 35 years were pretreated for 2 weeks with enteric-coated gar-
either slightly above or within therapeutic range, it is unlikely lic tablets (2000 mg of fresh garlic bulb), cranberry concen-
that they had a major effect on the anticoagulant properties of trate capsules (equal to 500 mg of cranberry juice concentrate
warfarin. Similar to the other 2 cases where the patients expe- or 57 g of cranberry fruit equivalent), or nothing at all. After
rienced subsequent increased INR results on more than one the pretreatment phase, a single oral dose of 25-mg racemic
occasion after consuming cranberry products, the DIPS rating warfarin (5 5 mg Coumadin tablets) was administered to
was found to be probable for this report (DIPS score ¼ 7 each subject. Participants continued taking the cranberry cap-
[þ1 þ 0 þ 0 þ 1 þ 1 þ 2 þ 0 þ 0 þ 1 þ 1]). sules and garlic tablets for 7 more days and then were subse-
quently crossed over to the other treatment group following a
Clinical trials. In the first clinical trial to investigate the poten- 2-week washout period. According to the results, there was
tial interaction between cranberry products and warfarin, Li not a significant pharmacokinetic interaction, as the plasma
et al conducted a randomized, placebo-controlled, double- concentration and unbound fraction of R- and S-warfarin was
blind, crossover study to evaluate the effects of cranberry unaffected by the concomitant administration of cranberry
juice on the anticoagulant effect of warfarin.30 Seven patients concentrate capsules. However, the cranberry capsules did
with atrial fibrillation, who had been stabilized on warfarin appear to enhance the pharmacodynamic effect of warfarin,
566 Journal of Pharmacy Practice 28(6)
as the area under the INR–time curve ([AUCINR] before war- 240 mL glasses of water or double-strength cranberry juice in
farin administration versus INR maximum after warfarin 15-minute intervals. A single dose of warfarin (Coumadin,
administration) increased by 30% when the capsules were Bristol-Myers Squibb, Princeton, New Jersey) 10 mg was
coadministered with warfarin when compared to warfarin administered with the third glass of cranberry juice. To prevent
1
treatment alone. Garlic did not appear to have a significant over-anticoagulation, 10 mg of vitamin K (Mephyton ; Merck,
effect on the pharmacokinetics or pharmacodynamics of Whitehouse station, New Jersey) was administered concomi-
warfarin. tantly with the dose of warfarin. According to the results, the
This randomized, double-blind trial evaluated the effects juice that inhibited warfarin’s metabolism in HLM had no
of cranberry juice on the pharmacokinetics and pharmacody- effect on the geometric mean AUC0-1 and terminal half-life
namics of warfarin.33 Thirty patients who were already of S/R warfarin in healthy participants. It was thereby con-
receiving warfarin for various indications were randomly cluded that cranberry juice does not significantly alter the
assigned to receive either 8 oz of cranberry juice or a metabolism of warfarin in humans.
matched placebo drink once daily for 2 weeks. Although According to the clinical trials reviewed, there does not
plasma concentrations for R-warfarin were found to be over- appear to be a significant pharmacokinetic or pharma-
all higher than those of S-warfarin, plasma levels of R- and codynamic interaction between warfarin and cranberry juice.
S-warfarin were not significantly different between the cran- Cranberry concentrate capsules, however, may increase antic-
berry juice group and the placebo group at any point during oagulant effect of warfarin and cause a pharmacodynamic
the study. Mean INR values were found to be similar interaction. Overall, the trials were limited by a relatively small
between the 2 study groups at all times during the study patient population. In the largest of these trials, 30 patients
except on treatment day 12, where the mean INR for the were evaluated. In each of the other trials, less than 20 patients
cranberry juice group was found to be significantly higher were included. With such small sample sizes, it is difficult to
than placebo (t ¼ 2.79; P < .02). It should be noted that at determine the significance of these studies. Another possible
the next measurement point (24 hours later), the mean INR limitation to the studies was the amount of cranberry juice uti-
levels for both groups were nearly the same. For unknown lized. Larger quantities of cranberry juice, similar to those sus-
reasons, 8 patients developed an INR >3.3 (4 in the cranberry pected to interact with warfarin in the majority of documented
juice group and 4 in the placebo group) and 1 patient in the case reports, could have been used to determine whether this
cranberry juice group developed a minimally reduced INR. potential drug–food interaction is dependent on the amount
The authors concluded that consuming 8 oz of cranberry of cranberry juice consumed. To better understand this poten-
juice daily does not significantly affect the metabolism or tial interaction, clinical trials involving more subjects who con-
anticoagulant effect of warfarin. sume larger amounts of cranberry juice are needed.
In this prospective, open-labeled study, the effects of ‘‘high-
dose’’ cranberry juice on the pharmacodynamics of warfarin
Patient considerations. Although the majority of evidence from
were examined.34 Ten patients on chronic warfarin therapy
clinical trial data suggests cranberry juice does not significantly
consumed 240 mL of cranberry juice twice daily for 7 days. interact with warfarin, the possibility of an interaction still
PTs and INR values were collected at baseline and on days
needs to be taken into consideration. As discussed, most of the
2, 6, and 8 of the study. The mean PTs from each study day
published case reports involved patients drinking larger
were then compared to the mean PT at baseline using repeated
amounts of cranberry juice than those used in clinical trials.
measures analysis of variance. There was no statistical differ-
Therefore, it is possible that a warfarin–cranberry juice interac-
ence noticed between the mean PT at baseline or at any time
tion exists but is dependent upon the amount of fruit juice
during the study, and it was concluded that a significant phar-
ingested. For now, patients who take warfarin should be
macodynamic interaction does not exist between high-dose
advised to consume cranberry products in moderation. A prac-
cranberry juice and warfarin. tical approach may be to limit the quantity of cranberry juice to
In the most recent trial to date, a prospective, open-labeled,
less than 8 oz, no more than 2 times per day, which is consistent
in vitro/in vivo study was conducted to determine the potential
with the amount of juice that failed to significantly alter the
inhibitory effects of cranberry juice on the hepatic CYP2C9-
antithrombotic effect of warfarin when studied. If abnormal
mediated metabolism of S-warfarin.35 Five different types of
INR readings occur and are thought to be related to the inges-
cranberry juice, purchased from 5 different vendors located
tion of cranberry juice or a cranberry product, patients should
in the United States, were first evaluated to determine their
be instructed to reduce the amount they drink/eat or discontinue
inhibitory effects on CYP2C9 activity (S-warfarin 7-hydroxy-
consuming cranberry products altogether.
lation) in human liver microsomes (HLM). Multiple bottles
of the cranberry juice found to be the most potent inhibitor
of CYP2C9 in HLM were then mixed together and converted
into a ‘‘double-strength’’ stock concentrate that was later used
Grapefruit (Citrus paradisi)
in the clinical phase of the trial. During the clinical study, 16 Grapefruit juice has been reported to inhibit the metabolism
healthy individuals (8 men and 8 nonpregnant women), aged of several medications metabolized by the CYP pathway.36
20 to 58 years, fasted overnight and then consumed 3 to It is speculated that the furanocoumarin, bergamottin, and the
Norwood et al 567
flavonoid, naringenin, are responsible for grapefruit’s CYP- In addition, the type of juice used could have also limited the
inhibiting properties. These 2 compounds usually demonstrate study. The investigators chose to use previously frozen grape-
a greater affinity for CYP isoforms located in the small intes- fruit juice; it is unknown if the freezing process itself can affect
tine as opposed to the liver and are believed to affect the the stability of grapefruit juice’s CYP inhibiting compounds. It
metabolism of several drugs that undergo first-pass metabo- is feasible that previously frozen and freshly prepared grape-
lism.37-39 Warfarin does not undergo significant first-pass fruit juice can have differing effects on drug metabolism.
metabolism and is not thought to be affected by the actions
of bergamottin and naringenin under normal conditions. How-
Patient considerations. As discussed, grapefruit juice is most
ever, there is evidence to suggest that bergamottin and narin-
likely to affect those medications that undergo first-pass
genin may inhibit both intestinal and hepatic CYP3A4 in an
metabolism. Drugs, such as warfarin, that have a relatively
exposure-dependent manner.40 Therefore, the metabolism of
high bioavailability and do not undergo significant first-
warfarin could be potentially impaired by large amounts of
pass metabolism, are not expected to interact with grapefruit
grapefruit juice.
juice. However, because increased amounts may potentially
affect warfarin’s metabolism, patients should be advised to
Case reports. A male patient in his 60s was found to have an
drink grapefruit juice in moderation. When studied by
INR of 6.29 after a year of being on warfarin therapy (50%
Sullivan et al,42 8 oz of grapefruit juice ingested 3 times
of his previous INR readings were within the target range of
daily was not shown to interact with warfarin, and patients
2-3).41 According to the report, the patient had been drinking
may want to limit their initial consumption to this amount.
nearly 50 oz (1479 mL) of grapefruit juice daily for 10 days
If INR levels increase and a warfarin–grapefruit juice interac-
prior to the supratherapeutic INR result. The patient was
tion is suspected, then it may be necessary to reduce the
instructed to hold warfarin therapy for 2 days, discontinue
amount of grapefruit juice consumed and/or monitor the INR
drinking grapefruit juice, and then resume his usual dose of
more frequently. Patients should also be encouraged to con-
warfarin. The majority of his next 10 INR values were within
sistently drink the same brand of grapefruit juice, as the
the target range of 2 to 3, with 3.24 reported as the highest INR
amount and potency of furanocoumarins and flavonoids may
value during that time. No bleeding episodes were reported.
differ among brands and types of grapefruit juice.36
Because warfarin is metabolized by several CYP isoenzymes
and there is scientific evidence that grapefruit juice may impair
the CYP pathway, the likelihood of a drug–food interaction
was increased in this report. Based on the DIPS, it was prob- Mango (Mangifera indica)
able that a warfarin–grapefruit juice occurred in this patient.
A clear mechanistic explanation for the potential interaction
(DIPS score ¼ 6 [þ0 þ 1 þ 1 þ 1 þ 1 þ 0 þ 0 þ 0 þ 1 þ 1]).
between warfarin and mango fruit has not yet been described;
however, it is speculated that vitamin A (retinol) may be
Clinical trials. To date, there has been only 1 clinical trial con-
responsible. Mangos contain varying amounts of vitamin A
ducted to evaluate the potential interaction between grapefruit
which may inhibit CYP2C19, an enzyme involved in the meta-
juice and warfarin.42 In this small open-label study, 9 men
bolism of warfarin’s R-isomer.43,44 It has been suggested that
between the ages of 55 and 75 years, stabilized on chronic war-
large doses of vitamin A can increase the anticoagulant effect
farin therapy, drank 8 oz (240 mL) of freshly prepared frozen
of warfarin.45
grapefruit juice (Smith’s Grapefruit Juice, Salt Lake City,
Utah) 3 times a day for 1 week to determine whether grapefruit
juice would affect their PTs. Compliance rates with drinking Case report. Thirteen male patients ranging from 56 to 85 years
the grapefruit juice ranged from 85.7% to 100%. There were of age experienced a 38% average increase in their INR from
no significant changes in the PT or INR during the study and baseline (P < .001) after eating approximately 1 to 6 mangos
it was concluded that the consumption of prepared, frozen daily, 2 days to 1 month prior to anticoagulation testing.46 Two
grapefruit juice did not change the PT in patients stabilized weeks after removing mango from their diet, there was a 17.7%
on warfarin therapy. average decrease in their INR from the previous result. There
Similar to most of the clinical trials that evaluated the effects were no bleeding episodes reported, and each of the men
of cranberry juice on warfarin, this study was limited by the denied alcohol use or changes in their health, diet, or medica-
small sample size and by the amount of fruit juice used. In the tion regimen. Two of the patients were rechallenged with the
case described by Bartle,41 the patient who experienced a same amount of mango that had previously consumed and their
supratherapeutic INR due to a possible grapefruit juice–war- INR levels subsequently increased. Two separate DIPS assess-
farin interaction consumed a larger amount of juice over a lon- ments were completed for this case report. For the 2 patients
ger period of time (50 oz daily for 10 days) compared to the who were rechallenged with mango, it was determined that a
amount used in the study (24 oz daily for 1 week). Thus, it is probable drug–food interaction occurred (DIPS score ¼ 6
possible that a significant interaction between grapefruit juice [þ0 þ 0 þ 0 þ 1 þ 1 þ 2 þ 0 þ 0 þ 1 þ 1]). In the other
and warfarin may not occur when the fruit juice is consumed 11 patients, use of the DIPS indicated a rating of possible
in small to moderate amounts but can when ingested in excess. (DIPS score ¼ 4 [þ0 þ 0 þ 0 þ 1 þ 1 þ 0 þ 0 þ 0 þ 1 þ 1]).
568 Journal of Pharmacy Practice 28(6)
Patient considerations. Unlike cranberry and grapefruit, mango pharmacokinetics and pharmacodynamics of warfarin must
products have not yet been evaluated in clinical trials to be evaluated in a controlled clinical trial. For now, it does not
determine whether they interact with warfarin. Even so, appear necessary to advise patients to limit their consumption
because several patients were documented as having an of pomegranate juice. However, because pomegranate juice
increase in their INR results after eating mangos, including may contain flavonoids and some other CYP-inhibiting com-
2 patients who were rechallenged, clinicians may want to pounds similar in nature to those found in grapefruits, providers
inquire about mango consumption as part of the dietary his- may want to ask about recent consumption in patients who
tory in areas where the fruit is popular and in those who experience fluctuations in their INR without any other known
experience elevated INR results for unknown reasons. If a causes.51 Patients known to drink large quantities of pomegra-
mango–warfarin interaction is suspected, then the amount nate juice should be encouraged to reduce their intake if a
of fruit consumed should be reduced. drug–food interaction is suspected.
for a parietooccipital brain infarct.57 Her INR was well con- use of over the counter products and/or alternative medicines,
trolled for the first 6 weeks of therapy but then suddenly existing infections, and poor dietary habits) that may have been
decreased from 2.5 to 1.7. The patient was taking the same dose partially, if not fully, responsible for the outcomes. In one case,
of warfarin and denied any medication changes or recent ill- the patient even ingested an extra dose of warfarin prior to
nesses. During this time, she admitted to eating almost 100 g being tested which would have obviously contributed to the
of avocado daily. She was advised to discontinue eating avo- abnormal INR result. Understanding the literature is further
cado and her INR returned to 2.5. Three months later, despite complicated by the paucity of adequate clinical trial data. Most
counseling, the patient again started eating avocado and her of the studies included a small patient population (participants
INR dropped to 1.7. Afterward, she completely eliminated avo- < 20) and used amounts of fruit juice that was less than that pur-
cado from her diet and her INR stabilized. It is well understood ported to interact with warfarin in case reports. In addition,
that foods containing vitamin K may potentially decrease the only 3 fruit products (cranberry juice, cranberry concentrate,
anticoagulant effect of warfarin. In addition, the likelihood of and grapefruit juice) have been studied with warfarin in a clin-
an interaction was increased in this case because the patient’s ical setting. Larger studies that use a variety of fruits, at varying
INR decreased in 2 separate occasions after eating avocado amounts, are needed to offer a clearer understanding of poten-
fruit. Therefore, using the DIPS, it was determined that a prob- tial warfarin-fruit interactions.
able drug–food interaction occurred in this patient (DIPS
score ¼ 8 [þ0 þ 1 þ 1 þ 1 þ 1 þ 2 þ 0 þ 0 þ 1 þ 1]).
The second report involves a 30-year-old female in her 22nd Conclusion
gestational week of her 6th pregnancy, who was anticoagulated In summary, there have been a number of case reports pub-
with heparin followed by warfarin due to a pulmonary embo- lished that suggest warfarin has the potential to interact with
lism.57 For 2 months, she was adequately anticoagulated with several fruit products, including cranberry juice and sauce,
an average INR of 2.7. After eating nearly 200 g of avocado grapefruit juice, mangos, pomegranate juice, avocados, and
on 2 consecutive evenings, her INR decreased to 1.6. She later papain (an enzyme found in the papaya fruit). It is difficult to
eliminated avocado from her diet and was able to maintain determine the relevance of these reports, however, as scientific
therapeutic INR levels. Similar to the other case involving avo- evidence is scarce. Thus far, only cranberry juice, cranberry
cado, the consumption of a vitamin K-containing food strength- concentrate, and grapefruit juice have been evaluated in a few
ened the possibility of an interaction in this report. Based on the human clinical trials to determine if they interact with warfarin.
DIPS, an interaction between avocado and warfarin was prob- Unfortunately, the majority of these studies have been limited
able in this case as well (DIPS score ¼ 6 [þ0 þ 1 þ 1 þ 1 þ 1 by small patient populations and the quantity of fruit juice used,
þ 0 þ 0 þ 0 þ 1 þ 1]). which has been considerably less than that reported to interact
with warfarin in case reports. To better ascertain potential
Patient considerations. In general, patients taking warfarin should warfarin-fruit interactions, more randomized controlled studies
be encouraged to eat vitamin K enriched foods in moderation involving a greater number of patients, a wider variety of fruits,
and to eat a consistent amount of these foods from week to and larger amounts of fruit product are warranted. Until further
week to avoid changes in the intensity of their anticoagulation information is available, clinicians may want to encourage
therapy. Clinicians may want to ask about the recent consump- patients to consume cranberry products and grapefruit juice
tion of avocado when obtaining a dietary history or when asses- in small to moderate quantities and to inquire about the recent
sing reasons for subtherapeutic INR results. Also, patients consumption of mangos, pomegranate juice, and avocados
should be discouraged from eating large amounts of avocado- when taking a dietary history or when assessing possible causes
based products, such as guacamole, as vitamin K ingestion can for INR instability.
be significantly higher with these foods.56
Declaration of Conflicting Interests
Limitations of the Literature The author(s) declared no potential conflicts of interest with respect to
the research, authorship, and/or publication of this article.
Interpreting the literature to determine which fruits may cause a
significant interaction with warfarin is a challenge. Individual
Funding
case reports currently makeup the bulk of evidence. Although
they can be helpful in informing clinicians of potential occur- The author(s) received no financial support for the research, author-
ship, and/or publication of this article.
rences, case reports in general are subjective and usually do not
contain enough information to establish a causative relation-
ship between a proposed interaction and outcome. In addition, References
they are often confounded by variables which, therefore, limit 1. Bartholow M. Top 200 drugs of 2012. Pharmacy Times. http://
their usefulness. This was found to be true as well for the cases www.pharmacytimes.com/publications/issue/2013/July2013/
reviewed in this article. The majority of reports contained very Top-200-Drugs-of-2012. Accessed May 1, 2014.
little patient-specific information leading up to the incidents 2. Coumadin (warfarin sodium oral tablets) package insert. Prince-
and were usually influenced by other factors (eg, concurrent ton, NJ: Bristol-Myers Squibb Company; 2014.
570 Journal of Pharmacy Practice 28(6)
3. Poller L. Progress in standardization in anticoagulant control. 21. Roberts D, Flanagan P. Case report: cranberry juice and warfarin.
Hematol Rev. 1987;1:225-241. Home Healthc Nurse. 2011;29(2):92-97.
4. Kirkwood TBL. Calibration of reference thromboplastins and 22. Hamann GL, Campbell JD, George CM. Warfarin-cranberry juice
standardisation of the prothrombin time ratio. Thromb Haemost. interaction. Ann Pharmacother. 2011;45(3):e17.
1983;49(3):238-244. 23. Haber SL, Cauthon KA, Raney EC. Cranberry and warfarin inter-
5. Guyatt GH, Akl EA, Crowther M, et al. Executive summary: action: a case report and review of the literature. Consult Pharm.
antithrombotic therapy and prevention of thrombosis, 9th ed: 2012;27(1):58-65.
American college of chest physicians evidence-based clinical 24. Hodek P, Trefil P, Stiborova M. Floavonoids-potent and versatile
practice guidelines. Chest. 2012;141(suppl 2):7S-47S. biologically active compounds interacting with cytochromes
6. Okey AB. Enzyme induction in the cytochrome P-450 system. P450. ChemBiol Interact. 2002;139(1):1-21.
Pharmacol Ther. 1990;45(2):241-298. 25. Kim E, Sy-Cordero A, Graf TN, et al. Isolation and identification
7. Rettie AE, Korzekwa KR, Kunze KL, et al. Hydroxylation of of intestinal CYP3A inhibitors from cranberry (Vaccinium
warfarin by human cDNA-expressed cytochrome P-450: a role for macrocarpon) using human intestinal microsomes. Planta Med.
P-4502C9 in the etiology of (S)-warfarin-drug interactions. Chem 2011;77(3):265-270.
Res Toxicol. 1992;5(1):54-59. 26. Greenblatt DJ, von Moltke LL. Interaction of warfarin with drugs,
8. Eble JM, West BD, Link KP. A comparison of the isomers of war- natural substances, and foods. J Clin Pharmacol. 2005;45(2):
farin. Biochem Pharmacol. 1966;15:1003-1006. 127-132.
9. Holbrook AM, Pereira JA, Labiris R, et al. Systematic overview 27. Duthie GC, Kylie JA, Jenkinson AM, et al. Increased salicy-
of warfarin and its drug and food interactions. Arch Intern Med. late concentrations in urine of human volunteers after con-
2005;165(10):1095-106. sumption of cranberry juice. J Agric Food Chem. 2005;
10. Committee on Safety of Medicines, Medicines and Healthcare Prod- 53(8):2897-900.
ucts Regulatory Agency. Interaction between warfarin and cran- 28. Wosilait WD. Theoretical analysis of the binding of salicylate by
berry juice: new advice. Curr Probl Pharmacovigil. 2004;30:10. human serum albumin: the relationship between the free and bound
11. Coumadin (warfarin sodium oral tablets) package insert. Prince- drug and therapeutic levels. Eur J ClinPharmacol. 1976;9(4):285-290.
ton, NJ: Bristol-Myers Squibb Company; 2005. 29. Renton KW. Cytochrome P450 regulation and drug biotransfor-
12. Warfarin. In: DRUGDEX1 System (Micromedex 2.0). Greenwood mation during inflammation and infection. Curr Drug Metab.
Village, CO: Truven Health Analytics; 2013. http://www.micro- 2004;5(3):235-243.
medexsolutions.com/micromedex2/librarian/ND_T/evidencexpert/ 30. Li Z, Seeram NP, Carpenter CL, et al. Cranberry does not affect
ND_PR/evidencexpert/CS/DE60B9/ND_AppProduct/evidencexpert/ prothrombin time in male subjects on warfarin. J Am Diet Assoc.
DUPLICATIONSHIELDSYNC/33F41D/ND_PG/evidencexpert/ND_ 2006;106(12):2057-2061.
B/evidencexpert/ND_P/evidencexpert/PFActionId/evidencexpert. 31. Lilja JJ, Find all citations by this author (defauOr filter your cur-
DisplayDrugdexDocument?docId=671285&contentSetId=100& rent search Backman JT, Find all citations by this author (defaul-
title=Warfarin+Sodium&servicesTitle=Warfarin+Sodium#3.5.2. tOr filter your current search Neuvonen PJ. Effects of Daily
G. Accessed November 13, 2013. Ingestion of Cranberry Juice on the Pharmacokinetics of
13. Horn JR, Hansten PD, Chan LN. Proposal for a new tool to eval- Warfarin, Tizanidine, and Midazolam–Probes off CYP2C9,
uate drug interaction cases. Ann Pharmacother. 2007;41(4): CYP1A2, and CYP3A4. Clin Pharm Ther. 2007;81(6):833-839.
674-680. 32. Mohammed Abdul MI, Jiang X, Williams KM, et al. Pharmaco-
14. Naranjo CA, Busto U, Sellers EM, et al. ‘‘A method for estimating dynamic interaction of warfarin with cranberry but not with garlic
the probability of adverse drug reactions’’. Clin Pharmacol Ther. in healthy subjects. Br J Pharmacol. 2008;154(8):1691-700.
1981;30(2):239-245. 33. Ansell J, McDonough M, Zhao Y, et al. The absence of an inter-
15. Suvarna R, Pirmohamed M, Henderson L. Possible interaction action between warfarin and cranberry juice: a randomized,
between warfarin and cranberry juice. BMJ. 2003;327(7429):1454. double-blind trial. J Clin Pharmacol. 2009;49(7):824-830.
16. Griffiths AP, Beddall A, Pegler S. Fatal haemopericardium and 34. Mellen CK, Ford M, Rindone JP. Effect of high-dose cranberry
gastrointestinal haemorrhage due to possible interaction of cran- juice on the pharmacodynamics of warfarin in patients. Br J Clin
berry juice with warfarin. J R Soc Promot Health. 2008;128(6): Pharmacol. 2010;70(1):139-142.
324-326. 35. Ngo N, Brantley SJ, Carrizosa DR, et al. The warfarin-cranberry
17. Grant P. Warfarin and cranberry juice: an interaction? J Heart juice interaction revisited: A systematic in vitro-in vivo evalua-
Valve Dis. 2004;13(1):25-26. tion. J Exp Pharmacol. 2010;2010(2):83-91.
18. Rindone JP, Murphy TW. Warfarin-cranberry juice interaction 36. Bailey DG, Malcolm J, Arnold O, et al. Grapefruit juice-drug
resulting in profound hypoprothrombinemia and bleeding. Am interactions 1998. Br J Clin Pharmacol. 2004;58:S831-S840.
J Ther. 2006;13(3):283-284. 37. Fuhr U, Klittich K, Staib AH. Inhibitory effect of grapefruit juice and
19. Paeng CH, Sprague M, Jackevicius CA. Interaction between war- its bitter principal, naringenin, on CYP1A2 dependent metabolism
farin and cranberry juice. Clin Ther. 2007;29(8):1730-1735. of caffeine in man. Br J Clin Pharmacol. 1993;35(4):431-436.
20. Mergenhagen KA, Sherman O. Elevated International Normal- 38. Ho PC, Saville DJ, Coville PF, et al. Content of CYP3A4 inhibi-
ized Ratio after concurrent ingestion of cranberry sauce and war- tors, naringin, naringenin and bergapten in grapefruit and grape-
farin. Am J Health Syst Pharm. 2008;65(22):2113-2116. fruit juice products. Pharm Acta Helv. 2000;74(4):379-385.
Norwood et al 571
39. Dresser GK, Bailey DG. The effects of fruit juices on drug dispo- 50. Jarvis S, Li C, Bogle RG. Possible interaction between
sition: a new model for drug interactions. Eur J Clin Invest. 2003; pomegranate juice and warfarin. Emerg Med J. 2010;27(1):
33(suppl 2):10-16. 74-75.
40. Veronese ML, Gillen LP, Burke JP, et al. Exposure-dependent 51. Kim H, Yoon YJ, Shon JH, et al. Inhibitory effects of fruit
inhibition of intestinal and hepatic CYP3A4 in vivo by grapefruit juice s on CYP3A activity. Drug Metab Dispos. 2006;34(4):
juice. J Clin Pharmacol. 2003;43(8):831-839. 521-523.
41. Bartle WR. Grapefruit juice might still be factor in warfarin 52. Shaw D, Leon C, Kolev S, et al. Traditional remedies and food
response. Am J Health Syst Pharm. 1999;56(7):676. supplements; a 5-year toxicological study (1991-1995). Drug Saf.
42. Sullivan DM, Ford MA, Boyden TW. Grapefruit juice and the 1997;17(5):342-356.
response to warfarin. Am J Health Syst Pharm. 1998;55(15): 53. Hewitt H, Wint Y, Talabere L, et al. The use of papaya on pres-
1581-1583. sure ulcers. Am J Nurs. 2002;102(12):73-77.
43. Yamazaki H, Shimada T. Effects of arachidonic acid, prostaglan- 54. Papain. In: AltMedDex1 System (Micromedex 2.0). Greenwood
dins, retinol, retinoic acid and cholecalciferol on xenobiotic oxi- Village, CO: Truven Health Analytics; 2013. http://www.micro
dations catalysed by human cytochrome P450 enzymes. medexsolutions.com/micromedex2/librarian/ND_T/evidencexpert/
Xenobiotica. 1999;29(3):231-241. ND_PR/evidencexpert/CS/54EFAD/ND_AppProduct/evidencexpe
44. Pennington JA. Bowes & Church’s Food Values of Portions Com- rt/DUPLICATIONSHIELDSYNC/9883D5/ND_PG/evidencexpert/
monly Used. 16th ed. Philadelphia: JB Lippincott, 1994. ND_B/evidencexpert/ND_P/evidencexpert/PFActionId/evidenc
45. Abramowicz M. Vitamin supplements. Med Lett Drugs Ther. expert.DisplayAltMedEvalDocument?docId=245&contentSetId=
1985;27:66-68. 60&topicId=cautionsSection&subtopicId=drugInteractionsSection.
46. Monterrey-Rodrı́guez J. Interaction between warfarin and mango Accessed December 20, 2013.
fruit. Ann Pharmacother. 2002;36(5):940-941. 55. Dismore ML, Haytowitz DB, Gebhardt SE, et al. Vitamin K con-
47. Gil MI, Tomás-Barberán FA, Hess-Pierce B, et al. Antioxidant activ- tent of nuts and fruits in the US diet. J Am Diet Assoc. 2003;
ity of pomegranate juice and its relationship with phenolic composi- 103(12):1650-1652.
tion and processing. J Agric Food Chem. 2000;48(10):4581-4589. 56. U.S. Department of Agriculture, Agricultural Research Service.
48. Hidaka M, Okumura M, Fujita K, et al. Effects of pomegranate 2013. USDA National Nutrient Database for Standard Reference,
juice on human cytochrome p450 3A (CYP3A) and carbamazepine Release 26. Nutrient Data Laboratory Home Page, http://www.
pharmacokinetics in rats. Drug Metab Dispos. 2005;33(5):644-648. ars.usda.gov/ba/bhnrc/ndl. Accessed February 17, 2014.
49. Komperda KE. Potential interaction between pomegranate juice 57. Blickstein D, Shaklai M, Inbal A. Warfarin antagonism by avo-
and warfarin. Pharmacotherapy. 2009;29(8):1002-1006. cado. Lancet. 1991;337(8746):914-915.