Bleeding risks in older patients on warfarin and taking antibiotics and antifungals
The study was conducted at the University of Texas Medical Branch, Galveston, by Baiullargeon J. et al. and published in the American Journal of Medicine in 2012.
The authors used enrollment, claims and pharmacy data in a 5% national sample of Medicare beneficiaries. Medicare Part A, which covers hospital expenses, begins automatically at age 65 years, whereas coverage for outpatient care (Part B) and prescription drugs (Part D) must be purchased. Claims from the 2007-2008 plan year for 5% of Medicare beneficiaries were used, including Medicare enrollment files, Medicare Provider Analysis and Review (MEDPAR) files, Outpatient Standard Analytic Files, Medicare Carrier files and Prescription Drug Event records.
Protocol
The study evaluated a cohort of 38,762 Medicare beneficiaries who were continuous users of warfarin. Presence and duration of warfarin use was examined by evaluating prescription data from the Medicare Part D dataset. Continuous warfarin users were followed from January 1, 2008, until hospitalization for a bleeding event or the end of study period (December 31, 2008), whichever occurred first. Reasons for warfarin use included atrial fibrillation, stroke, presence of prosthetic heart valve and venous thromboembolism.
Definition of cases and controls
Cases were defined as patients who experienced a bleeding event requiring hospitalization at any time in 2008, based on International Classification of Diseases, ninth revision codes (ICD-9). Types of bleeding identified were gastrointestinal, non-gastrointestinal, intracranial and general warfarin toxicity. The event date was defined as the date of hospital admission. Controls were matched with cases on event date, indication for warfarin use, age, sex and race/ethnicity.
Definition of antibiotic exposure
Antibiotic exposure was determined by assessing the number of days in a given prescription period that followed the initial prescription date in the Medicare Part D dataset. Patients whose most recent prescription period for any antibiotic agent overlapped by a least one day with the 15-day period before the event date were defined as exposed. Among patients with antibiotic exposure, the authors categorized the time between the initiation of antibiotic agents and the event date as 0-15 days, 16-60 days and more than 60 days before the event date. The antibiotics were categorized as "all antibiotic medication" and assigned to the following classes of antibiotics: cephalosporins, penicillins, cotrimoxazole, macrolides and quinolones. Antifungals were categorized as azole antifungals.
Medications as potential confounders
The authors also examined patients' use of potentially confounding medications known to interact with warfarin. A prescription for any potential confounding drugs that included at least 1 day in the 15 days before the event date was defined as presence of confounding drugs. The following classes of drugs were examined: antidepressants, antiplatelet drugs, corticosteroids and selected inhibitors of warfarin hepatic metabolism.
Statistical Analysis
Standard regression analysis was used to calculate odds ratios and 95% confidence intervals for the risk of bleeding associated with prior exposure to antibiotic medications.
Study results
The authors identified 38,762 patients as continuous warfarin users.
During 2008, 1136 of those (2.9%) were hospitalized with a primary diagnosis of bleeding. Of those, 798 met the definition for a case.
Continuous warfarin users exposed to any antibiotic agent were twice as likely (odds ratio of 2.01) to experience a bleeding event that required hospitalization compared to those who were not exposed to an antibiotic.
Assessment of type of bleeding events showed that antibiotic users were two and a half times more likely to experience non-gastrointestinal bleeding and about twice as likely to experience a gastrointestinal bleed.
The authors examined, among all patients exposed to antibiotics, as to whether recency of prescription initiation was associated with major bleeding. Among all exposed patients, those whose prescription began in the 0-15 days or 16-60 days before the event date were more likely to have been hospitalized for bleeding compared with control patients who never took antibiotics.
Exposed patients whose antibiotic prescription began greater than 60 days before the event date did not have a statistically significant increased risk for bleeding that required hospitalization compared with the control patients.
Patients treated with azole antifungals were over 4 times more likely to experience bleeding; treatment with macrolides, a little less than twice as likely to experience bleeding; quinolones was 1.7 times as likely; cotrimoxazole was 2.7 times as likely; cephalosporins 2.5 times as likely; and penicillins about twice as likely to experience bleeding.
Of the confounder medications, the serotonin selective reuptake inhibitors and corticosteroids were associated with statistically significant increased risk of bleeding.
Discussion points
Exposure to any antibiotic agent was associated with a 2-fold increased risk of bleeding that required hospitalization among older continuous warfarin users.
When broken down by class, all 5 specific antibiotic drug classes were associated with an increased risk of bleeding.
Patients who were prescribed azole antifungals and cotrimoxazole had the highest risks of hospitalization for bleeding. According to the authors, this finding was consistent with previous reports.
The increased risk associated with antifungals and cotrimoxazole was attributed to their inhibitory effects on CYP2C9 - one of the liver enzymes that metabolize warfarin.
Summary of this study
Numerous studies have reported that interactions between warfarin and antibiotics may result in an increased international normal ratio (INR). The authors suggested that frequent monitoring of INR has been recommended for patients who are concurrently taking warfarin and antibiotic agents. Since inhibition of vitamin K synthesis by alteration of gut flora or inhibition of cytochrome P450 enzymes can lead to an increased INR and increased bleeding within a 1 to 2 week period, a suggested strategy is to monitor INR one week after initiating antibiotic therapy.
The full report can be found at Baiullargeon J. et al. Concurrent use of warfarin and antibiotics and the risk of bleeding in older adults. The American Journal of Medicine 2012; 125:183-89.
Interaction exercise using the Lexi-Interact™ application
Many of the antibiotics and antifungals, shown in this study to increase bleeding risks in older patients on warfarin, are those used in dentistry. These include fluconazole, ketoconazole, miconazole, cephalexin, cefaclor, azithromycin, clarithromycin, amoxicillin, penicillin VK ciprofloxacin, levofloxacin and moxifloxacin.
The Lexi-Interact application is a useful tool to assess the potential for bleeding risks in warfarin patients when prescribing antibiotics or antifungals, and to assess the management of these patients.
An analysis was performed by this author for two sample antibiotics using Lexi-Interact in order to evaluate the potential for bleeding in warfarin patients. Below is a summary of the database information for amoxicillin and clindamycin with warfarin as the interacting drug. Amoxicillin does interact with warfarin and clindamycin does not interact with warfarin.
Amoxicillin - Warfarin interaction analysis from Lexi-Interact database
Title: Vitamin K Antagonists (warfarin) / Penicillins (amoxicillin)
Summary: Penicillins (amoxicillin) may enhance the anticoagulant effect of Vitamin K Antagonists (warfarin).
Interaction Risk Rating is listed as C: The severity of the interaction is listed as moderate. The data demonstrate that the specified agents may interact with each other in a clinically significant manner. The benefits of concomitant use of these two medications usually outweigh the risks. An appropriate monitoring plan should be implemented to identify potential negative effects. Dosage adjustments of one or both agents may be needed in a minority of patients.
Patient Management: Monitor anticoagulation status more closely in vitamin K antagonist-treated patients both during concurrent use of a penicillin antibiotic and for at least several days after cessation of the antibiotic. Similarly, advise patients to pay particularly close attention to any signs or symptoms of bleeding. Penicillin antibiotics with a broader spectrum of activity (especially including those with a concurrent beta-lactamase inhibitor) may be most likely to cause a significant interaction.
Clindamycin - Warfarin interaction analysis from Lexi-Interact database
No interactions of Risk Level A or greater identified.
Risk level A = No Known Interaction. Data have not demonstrated either pharmacodynamic or pharmacokinetic interactions between the specified agents.
