Drug Interactions between isoniazid / pyrazinamide / rifampin and warfarin

MONITOR: Vitamin K may antagonize the hypoprothrombinemic effect of oral anticoagulants. Vitamin K is a cofactor in the synthesis of blood clotting factors that are inhibited by oral anticoagulants, thus intake of vitamin K through supplements or diet can reverse the action of oral anticoagulants. Resistance to oral anticoagulants has been associated with consumption of foods or enteral feedings high in vitamin K content. Likewise, a reduction of vitamin K intake following stabilization of anticoagulant therapy may result in elevation of the INR and bleeding complications. Foods rich in vitamin K include beef liver, broccoli, Brussels sprouts, cabbage, collard greens, endive, kale, lettuce, mustard greens, parsley, soy beans, spinach, Swiss chard, turnip greens, watercress, and other green leafy vegetables. Moderate to high levels of vitamin K are also found in other foods such as asparagus, avocados, dill pickles, green peas, green tea, canola oil, margarine, mayonnaise, olive oil, and soybean oil. Snack foods containing the fat substitute, olestra, are fortified with 80 mcg of vitamin K per each one ounce serving so as to offset any depletion of vitamin K that may occur due to olestra interference with its absorption. Whether these foods can alter the effect of oral anticoagulants has not been extensively studied. One small study found that moderate consumption (1.5 servings/day) does not significantly affect the INR after one week in patients receiving long-term anticoagulation.

Consumption of large amounts of mango fruit has been associated with enhanced effects of warfarin. The exact mechanism of interaction is unknown but may be related to the vitamin A content, which may inhibit metabolism of warfarin. In one report, thirteen patients with an average INR increase of 38% reportedly had consumed one to six mangos daily 2 to 30 days prior to their appointment. The average INR decreased by 17.7% after discontinuation of mango ingestion for 2 weeks. Rechallenge in two patients appeared to confirm the interaction.

Limited data also suggest a potential interaction between warfarin and cranberry juice resulting in changes in the INR and/or bleeding complications. The mechanism is unknown but may involve alterations in warfarin metabolism induced by flavonoids contained in cranberry juice. At least a dozen reports of suspected interaction have been filed with the Committee on Safety of Medicines in the U.K. since 1999, including one fatality. In the fatal case, the patient's INR increased dramatically (greater than 50) six weeks after he started drinking cranberry juice, and he died from gastrointestinal and pericardial hemorrhage. However, the patient was also taking cephalexin for a chest infection and had not eaten for two weeks prior to hospitalization, which may have been contributing factors. Other cases involved less dramatic increases or instabilities in INR following cranberry juice consumption, and a decrease was reported in one, although details are generally lacking. In a rare published report, a 71-year-old patient developed hemoptysis, hematochezia, and shortness of breath two weeks after he started drinking 24 ounces of cranberry juice a day. Laboratory test results on admission revealed a decrease in hemoglobin, an INR greater than 18, and prothrombin time exceeding 120 seconds. The patient recovered after warfarin doses were withheld for several days and he was given packed red blood cells, fresh-frozen plasma, and subcutaneous vitamin K. It is not known if variations in the constituents of different brands of cranberry juice may affect the potential for drug interactions.

There have been several case reports in the medical literature of patients consuming grapefruit, grapefruit juice, or grapefruit seed extract who experienced increases in INR. R(+) warfarin, the less active of the two enantiomers of warfarin, is partially metabolized by CYP450 3A4. Depending on brand, concentration, dose and preparation, grapefruit juice may be considered a moderate to strong inhibitor of CYP450 3A4, thus coadministration with warfarin may decrease the clearance of R(+) warfarin. However, the clinical significance of this effect has not been established. A pharmacokinetic study found no effect on the PT or INR values of nine warfarin patients given 8 oz of grapefruit juice three times a day for one week.

A patient who was stabilized on warfarin developed a large hematoma in her calf in association with an elevated INR of 14 following consumption of approximately 3 liters of pomegranate juice in the week prior to admission. In vitro data suggest that pomegranate juice can inhibit CYP450 2C9, the isoenzyme responsible for the metabolic clearance of the biologically more active S(-) enantiomer of warfarin. In rats, pomegranate juice has also been shown to inhibit intestinal CYP450 3A4, the isoenzyme that contributes to the metabolism of R(+) warfarin.

Black currant juice and black currant seed oil may theoretically increase the risk of bleeding or bruising if used in combination with anticoagulants. The proposed mechanism is the antiplatelet effects of the gamma-linolenic acid constituent in black currants.

Soy protein in the form of soy milk was thought to be responsible for a case of possible warfarin antagonism in an elderly male stabilized on warfarin. The exact mechanism of interaction is unknown, as soy milk contains only trace amounts of vitamin K. Subtherapeutic INR values were observed approximately 4 weeks after the patient began consuming soy milk daily for the treatment of hypertriglyceridemia. No other changes in diet or medications were noted during this time. The patient's INR returned to normal following discontinuation of the soy milk with no other intervention.

An interaction with chewing tobacco was suspected in a case of warfarin therapy failure in a young male who was treated with up to 25 to 30 mg/day for 4.5 years. The inability to achieve adequate INR values led to eventual discontinuation of the chewing tobacco, which resulted in an INR increase from 1.1 to 2.3 in six days. The authors attributed the interaction to the relatively high vitamin K content in smokeless tobacco.

MANAGEMENT: Intake of vitamin K through supplements or diet should not vary significantly during oral anticoagulant therapy. The diet in general should remain consistent, as other foods containing little or no vitamin K such as mangos and soy milk have been reported to interact with warfarin. Some experts recommend that continuous enteral nutrition should be interrupted for one hour before and one hour after administration of the anticoagulant dose and that enteral formulas containing soy protein should be avoided. Patients should also consider avoiding or limiting the consumption of cranberry juice or other cranberry formulations (e.g., encapsulated dried cranberry powder), pomegranate juice, black currant juice, and black currant seed oil.

GENERALLY AVOID: Concurrent use of rifampin in patients who ingest alcohol daily may result in an increased incidence of hepatotoxicity. The increase in hepatotoxicity may be due to an additive risk as both alcohol and rifampin are individually associated with this adverse reaction. However, the exact mechanism has not been established.

ADJUST DOSING INTERVAL: Administration with food may reduce oral rifampin absorption, increasing the risk of therapeutic failure or resistance. In a randomized, four-period crossover phase I study of 14 healthy male and female volunteers, the pharmacokinetics of single dose rifampin 600 mg were evaluated under fasting conditions and with a high-fat meal. Researchers observed that administration of rifampin with a high-fat meal reduced rifampin peak plasma concentration (Cmax) by 36%, nearly doubled the time to reach peak plasma concentration (Tmax) but reduced overall exposure (AUC) by only 6%.

MANAGEMENT: The manufacturer of oral forms of rifampin recommends administration on an empty stomach, 30 minutes before or 2 hours after meals. Patients should be encouraged to avoid alcohol or strictly limit their intake. Patients who use alcohol and rifampin concurrently or have a history of alcohol use disorder may require additional monitoring of their liver function during treatment with rifampin.

GENERALLY AVOID: Concurrent use of isoniazid (INH) in patients who ingest alcohol daily may result in an increased incidence of both hepatotoxicity and peripheral neuropathy. The increase in hepatotoxicity may be due to an additive risk as both alcohol and INH are individually associated with this adverse reaction. INH-associated hepatotoxicity is believed to be due to an accumulation of toxic metabolites and may also be partly immune mediated, though the exact mechanisms are not universally agreed upon. INH is metabolized by N-acetyltransferase and CYP450 2E1. The rate of acetylation is genetically determined and generally classified as slow or rapid. Slow acetylators have been identified by some studies as having a higher risk of hepatotoxicity; therefore, this interaction may be more significant for patients who fall into this category. Other studies have postulated that alcohol-mediated CYP450 2E1 induction may play a role, as this isoenzyme is involved in INH metabolism and may be responsible for producing hepatotoxic metabolites. However, available literature is conflicting. The labeling for some INH products lists daily alcohol use or chronic alcoholism as a risk factor for hepatitis, but not all studies have found a significant association between alcohol use and INH-induced hepatotoxicity. Additionally, INH and alcohol are both associated with pyridoxine (B6) deficiency, which may increase the risk of peripheral neuropathy.

GENERALLY AVOID: Concomitant administration of isoniazid (INH) with foods containing tyramine and/or histamine may increase the risk of symptoms relating to tyramine- and/or histamine toxicity (e.g., headache, diaphoresis, flushing, palpitations, and hypotension). The proposed mechanism is INH-mediated inhibition of monoamine oxidase (MAO) and diamine oxidase (DAO), enzymes responsible for the metabolism of tyramine and histamine, respectively. Some authors have suggested that the reactions observed are mainly due to INH's effects on DAO instead of MAO or the amounts of histamine instead of tyramine present in the food. A Japanese case report recorded an example in 8 out of 25 patients on the tuberculosis ward who developed an accidental histamine poisoning after ingesting a fish paste (saury). Patients developed allergy-like symptoms, which started between 20 minutes and 2 hours after ingesting the food. A high-level of histamine (32 mg/100 g of fish) was confirmed in the saury paste and all 8 patients were both on INH and had reduced MAO concentrations. The 17 remaining patients were not on INH (n=5) or reported not eating the saury paste (n=12).

ADJUST DOSING INTERVAL: Administration with food significantly reduces oral isoniazid (INH) absorption, increasing the risk of therapeutic failure or resistance. The mechanism is unknown. Pharmacokinetic studies completed in both healthy volunteers (n=14) and tuberculosis patients (n=20 treatment-naive patients during days 1 to 3 of treatment) have resulted in almost doubling the time to reach INH's maximum concentration (tmax) and a reduction in isoniazid's maximum concentration (Cmax) of 42%-51% in patients who consumed high-fat or high-carbohydrate meals prior to INH treatment.

MANAGEMENT: The manufacturer of oral forms of isoniazid (INH) recommends administration on an empty stomach (i.e., 30 minutes before or 2 hours after meals). Patients should be encouraged to avoid alcohol or strictly limit their intake. Patients who use alcohol and INH concurrently or have a history of alcohol use disorder may require additional monitoring of their liver function during treatment with INH. Concomitant pyridoxine (B6) administration is also recommended to reduce the risk of peripheral neuropathy, with some authorities suggesting a dose of at least 10 mg/day. Patients should be advised to avoid foods containing tyramine (e.g., aged cheese, cured meats such as sausages and salami, fava beans, sauerkraut, soy sauce, beer, or red wine) or histamine (e.g., skipjack, tuna, mackerel, salmon) during treatment with isoniazid. Consultation of product labeling for combination products containing isoniazid and/or relevant guidelines may be helpful for more specific recommendations.

MONITOR: Enhanced hypoprothrombinemic response to warfarin has been reported in patients with acute alcohol intoxication and/or liver disease. The proposed mechanisms are inhibition of warfarin metabolism and decreased synthesis of clotting factors. Binge drinking may exacerbate liver impairment and its metabolic ability in patients with liver dysfunction. The risk of bleeding may be increased. Conversely, reductions in INR/PT have also been reported in chronic alcoholics with liver disease. The proposed mechanism is that continual drinking of large amounts of alcohol induces the hepatic metabolism of anticoagulants. Effects are highly variable and significant INR/PT fluctuations are possible.

MANAGEMENT: Patients taking oral anticoagulants should be counseled to avoid large amounts of ethanol, but moderate consumption (one to two drinks per day) are not likely to affect the response to the anticoagulant in patients with normal liver function. Frequent INR/PT monitoring is recommended, especially if alcohol intake changes considerably. It may be advisable to avoid oral anticoagulant therapy in patients with uncontrollable drinking problems. Patients should be advised to promptly report any signs of bleeding to their doctor, including pain, swelling, headache, dizziness, weakness, prolonged bleeding from cuts, increased menstrual flow, nosebleeds, bleeding of gums from brushing, unusual bleeding or bruising, red or brown urine, or red or black stools.

MONITOR: Multivitamin preparations containing vitamin K may antagonize the hypoprothrombinemic effect of oral anticoagulants in some patients. Vitamin K1 in its active, reduced form serves as a cofactor in the generation of functional clotting factors, during which it becomes oxidized. It is reactivated in a process that is inhibited by oral anticoagulants, thus intake of additional vitamin K through supplements or diet can reverse the action of oral anticoagulants. Although the amount of vitamin K in over-the-counter multivitamin preparations is generally well below the dose thought to affect anticoagulation, there have been isolated case reports of patients stabilized on warfarin whose INR decreased following initiation of a multivitamin supplement and returned to therapeutic levels upon cessation of the multivitamin. Increases in warfarin dosage were required in some cases when the multivitamin was continued. One patient whose warfarin dosage was increased developed a subcapsular hematoma in her right kidney two weeks after she discontinued the multivitamin without informing her physician. Her INR was 13.2 and she was treated with vitamin K and fresh frozen plasma. It is possible that patients with low vitamin K status may be particularly susceptible to the interaction. Investigators have shown that vitamin K deficiency can cause an oversensitivity to even small increases in vitamin K intake. In one study where warfarin-stabilized patients were given a multivitamin tablet containing 25 mcg of vitamin K1 daily for 4 weeks, subtherapeutic INRs occurred in 9 of 9 patients with low vitamin K1 levels (<1.5 mcg/L) and only 1 of 7 patients with normal vitamin K1 levels (>4.5 mcg/L). INR decreased by a median of 0.51 and warfarin dosage had to be increased by 5.3% in patients with low vitamin K1 levels, whereas INR and warfarin dosage did not change significantly in patients with normal vitamin K1 levels. The prevalence of vitamin K deficiency may be small, but significant in the anticoagulated population. In a survey of 179 consecutive ambulatory patients on stable warfarin therapy attending an anticoagulation clinic, 22 (12.3%) were found to have vitamin K1 deficiency (<0.1 ng/mL).

MANAGEMENT: The potential for multivitamin supplements containing even low levels of vitamin K to affect anticoagulation should be recognized. In particular, elderly and/or malnourished patients may require more frequent monitoring of INR following the initiation or discontinuation of a multivitamin supplement, and the anticoagulant dosage adjusted as necessary.

MONITOR: Smoking cessation may lead to elevated plasma concentrations and enhanced pharmacologic effects of drugs that are substrates of CYP450 1A2 (and possibly CYP450 1A1) and/or certain drugs with a narrow therapeutic index (e.g., flecainide, pentazocine). One proposed mechanism is related to the loss of CYP450 1A2 and 1A1 induction by polycyclic aromatic hydrocarbons in tobacco smoke; when smoking cessation agents are initiated and smoking stops, the metabolism of certain drugs may decrease leading to increased plasma concentrations. The mechanism by which smoking cessation affects narrow therapeutic index drugs that are not known substrates of CYP450 1A2 or 1A1 is unknown. The clinical significance of this interaction is unknown as clinical data are lacking.

MANAGEMENT: Until more information is available, caution is advisable if smoking cessation agents are used concomitantly with drugs that are substrates of CYP450 1A2 or 1A1 and/or those with a narrow therapeutic range. Patients receiving smoking cessation agents may require periodic dose adjustments and closer clinical and laboratory monitoring of medications that are substrates of CYP450 1A2 or 1A1.