Clinicians commonly prescribe nonsteroidal anti-inflammatory drugs (NSAIDs) to treat acute and chronic pain and especially inflammation. NSAIDs exert their anti-inflammatory effect by inhibiting cyclooxygenase 2 (COX-2) enzymes, which inhibit the production of various prostaglandins. The selectivity of a particular NSAID for COX-1 and/or COX-2 enzymes leads to variation in their activity and toxicity. Nonselective NSAIDs comprise 3 categories: acetylated salicylates (aspirin), nonacetylated salicylates (choline magnesium trisalicylate, diflunisal, and salsalate), and nonsalicylates (etodolac, ibuprofen, indomethacin, meclofenamate, meloxicam, nabumetone, naproxen, etc). Certain COX-2 selective NSAIDs are also known as the coxibs, such as celecoxib and parecoxib (available in Europe). Some noncoxib NSAIDs have also been shown to be COX-2 selective. These include etodolac, meloxicam, and nabumetone.1

Although investigators have completed considerable research on drug–drug interactions between anticoagulants and NSAIDs, studies that focus specifically on diflunisal and warfarin are limited. The following is a brief review of these agents and an analysis of the theoretical mechanism of their interaction.

Warfarin, a vitamin K antagonist, is FDA approved for the prophylaxis and treatment of venous thrombosis and thromboembolic complications associated with atrial fibrillation and valve replacements.2  Specifically, warfarin works by inhibiting the enzyme complex vitamin K epoxide reductase, which is necessary to reform vitamin K in the normal clotting cycle. By impeding this process, warfarin prevents the generation of vitamin K–dependent clotting factors II, VII, IX, and X, resulting in disruption of the coagulation cascade.

Diflunisal is a prescription-only, nonselective NSAID with a salicylate moiety, approved to treat mild to moderate pain, osteoarthritis, and rheumatoid arthritis. Although it is a difluorophenyl derivative of salicylic acid, differing from aspirin structurally by only 2 substituents, it is not metabolized to salicylic acid.3 Because of its lack of selectivity, diflunisal reversibly inhibits both COX-1 and COX-2 enzymes. Inhibition of COX-1, an enzyme that catalyzes production of prostaglandins specific to the platelet aggregation cascade, can lead to obvious implications on bleeding risk. However, at usual analgesic doses, nonacetylated salicylates, like diflunisal, typically do not have a significant effect on platelet function and are therefore presumably associated with decreased gastrointestinal (GI) bleeding compared with other nonselective NSAIDs. Data have shown that greater COX-2 selectivity also has reduced GI toxicity. However, the results of several studies have demonstrated that coxibs have an elevated risk for thromboembolism over traditional NSAIDs.4 Notwithstanding, aspirin has a greater effect on platelet function because of irreversible inhibition of the COX-1 enzyme. The more concerning characteristic of diflunisal in the context of warfarin is its protein-binding affinity. Normally, warfarin is 99% protein bound in blood.2 Diflunisal, like aspirin and other NSAIDs, competitively displaces warfarin from its plasma protein binding site, increasing the amount of free anticoagulant in plasma.3 This could theoretically increase prothrombin times and put the patient at increased risk of bleeding.

One of the few studies completed on this specific interaction was published in 1980 and enrolled just 5 participants. The authors found that the mean percentage of unbound warfarin increased directly with the mean plasma concentration of diflunisal with a coefficient of correlation of 0.951 (P <.001). This is of particular concern when considering that, because of its long half-life, a loading dose of 500 mg is recommended when initiating diflunisal.3 However, although data analysis showed a 30% increase in unbound warfarin concentrations with concomitant diflunisal administration, the investigators also observed a 28% drop in the total (systemic) warfarin concentration. This suggests that compensatory clearance was offsetting the increase in free warfarin. Importantly, the results of this study did not demonstrate an increase in anticoagulant effect while diflunisal was being coadministered, but they did show a decrease in anticoagulant effect when diflunisal was withdrawn.5

Additional studies examining concomitant warfarin and diflunisal are needed to thoroughly understand the implications on physiological and serological factors and resultant health outcomes. Although theoretically COX-2 selective NSAIDs may be safer from a bleeding perspective compared with nonselective NSAIDs when considering GI toxicity, data demonstrate a similar rate of hospitalization due to GI bleeding when either NSAID is used concurrently with warfarin.6 For now, if concomitant therapy cannot be avoided, clinicians should monitor international normalized ratios more frequently when discontinuing or starting any NSAID in tandem with warfarin. And as always with warfarin therapy, consistency is key.
 
Stephen M. Chamberlin, Sharon A. Cherian, and Amy Lee are PharmD candidates at Albany College of Pharmacy and Health Sciences (ACPHS) in New York.

Amy T. Murdico, PharmD, BCPS, is the associate chief of pharmacy services, the PGY-1 pharmacy residency director, and the PGY-2 pain and palliative care pharmacy residency coordinator at Albany Stratton VA Medical Center in New York. She coordinates and precepts in the VA Learning Opportunities Residency intern program and is an adjunct advanced pharmacy practice experience preceptor for ACPHS and Western New England University in Springfield, Massachusetts.

Jeffrey Fudin, PharmD, FCCP, FASHP, FFSMB, is CEO and founder of Remitigate LLC and the owner and managing editor of paindr.com. He is also an adjunct associate professor at Western New England University College of Pharmacy and Health Sciences, an adjunct associate professor of pharmacy practice and pain management at ACPHS, and the director of the PGY-2 pain residency at Albany Stratton VA Medical Center.


REFERENCES
  1. Yang M, Wang HT, Zhao M, et al. Network meta-analysis comparing relatively selective COX-2 inhibitors versus coxibs for the prevention of NSAID-induced gastrointestinal injury. Medicine (Baltimore). 2015;94(40):e1592. doi: 10.1097/MD.0000000000001592.
  2. Coumadin [prescribing information]. Princeton, NJ: Bristol-Myers Squibb Co; 2017. packageinserts.bms.com/pi/pi_coumadin.pdf. Accessed July 23, 2019.
  3. Dolobid [prescribing information]. Whitehouse Station, NJ: Merck & Co Inc; 2007. www.accessdata.fda.gov/drugsatfda_docs/label/2007/018445s058lbl.pdf. Accessed July 23, 2019.
  4. Varga Z, Sabzwari S, Vargova V. Cardiovascular risk of nonsteroidal anti-inflammatory drugs: an under-recognized public health issue. Cureus. 2017;9(4):e1144. doi: 10.7759/cureus.1144.
  5. Serlin MJ, Mossman S, Sibeon RG, Tempero KF, Breckenridge AM. Interaction between diflunisal and warfarin. Clin Pharmacol Ther. 1980;28(4):493-498. doi: 10.1038/clpt.1980.193.
  6. Battistella M, Mamdami MM, Juurlink DN, Rabeneck L, Laupacis A. Risk of upper gastrointestinal hemorrhage in warfarin users treated with nonselective NSAIDs or COX-2 inhibitors. Arch Intern Med. 2005;165(2):189-192. doi: 10.1001/archinte.165.2.189.