In our patient case, treatment with medical cannabis did not significantly impact warfarin therapy and INR levels remained stable. This observed effect is contrary to other reports that suggest cannabis may interact with warfarin therapy and lead to increased INR levels (Damkier et al. 2019; Yamreudeewong et al. 2009; Hsu and Painter 2019; Grayson et al. 2017; Brown et al. 2021). THC is the primary psychoactive compound present in cannabis. CBD, another major cannabinoid compound, is believed to contribute to cannabis’ therapeutic effects (Health Canada 2018; Pertwee 2014). Available dosage forms for cannabis include capsules, oils, tinctures, lozenges, edibles, topicals, rectal suppositories, and oromucosal spray (MacCallum and Russo 2018). Medical cannabis products may be prepared for oral, oromucosal, nasal, transdermal, and rectal administration (MacCallum and Russo 2018). Inhalation of the aerosols from vaporization (i.e., “vaping”) or combustion (i.e., “smoking”) are also common methods of administration (MacCallum and Russo 2018).

Our search revealed 5 relevant articles describing INR elevations in patients on warfarin who were also using cannabis products. Two case reports revealed INR elevation with smoked cannabis for recreational use (Damkier et al. 2019; Yamreudeewong et al. 2009). A third case report also showed INR elevation with primarily edible cannabis and occasional smoked cannabis use that was prescribed for anxiety and attention deficit hyperactivity disorder (Hsu and Painter 2019). A fourth case report describes INR elevation in a patient on warfarin receiving two medical cannabis sublingual oil products at a total daily dose of 14.7 mg THC and 10.3 mg CBD (Brown et al. 2021).

In addition to previously published case reports, information on the FDA-approved drug, CBD oral solution (Epidiolex®), may be helpful in predicting the effects medical cannabis may have on metabolism of concurrent drugs. The manufacturer recommends considering dose reductions of CYP2C9 substrates, such as warfarin, in patients treated with CBD oral solution. The fifth case report found that CBD oral solution, administered at a dose starting at 5 mg/kg/day and titrated up to 35 mg/kg/day, did impact warfarin therapy resulting in an elevated INR (Grayson et al. 2017).

Our patient’s medical cannabis regimen delivered between 0.064 and 0.072 mg/kg/day of CBD and did not elevate his INR. In vitro studies suggest THC and CBD both are capable of inhibiting CYP2C9 and may increase warfarin's effect in a dose-dependent manner (Health Canada 2018; Yamaori et al. 2012). We theorize that the THC and CBD amounts our patient was exposed to were lower than the amounts required to inhibit CYP2C9’s warfarin metabolism. Our theory is supported by the low ratios of the estimated maximum serum concentrations (C_max) of CBD and THC and the in vitro inhibitory concentrations (K_i). According to Kiyomi et al., a C_max/K_i ratio < 0.1 is considered to have a low risk of causing a clinically observed drug-drug interaction (Kiyomi et al. 2004). The C_max/K_i ratios for CBD and THC for this patient were 4.9 × 10⁻⁴ and 7.3 × 10⁻⁴, respectively. Because we did not have the patient’s serum CBD or THC levels, we estimated CBD and THC C_max by using dosing data reported by Miller et al. to find the best-fit line and its equation for CBD (Miller et al. 2018). Table 2 shows the CBD doses, C_max, and the equation and R² value for the best-fit trend line. Using the estimated C_max, 0.87 ng/mL (by entering the patient’s daily CBD dose into the best-fit line equation), the K_i determined by Yamaori (5.6 μM), and CBD’s molecular weight (314.47 ng/nanomoles), we calculated the ratio as follows:

To estimate our patient’s THC C_max, we utilized the C_max and THC dosing data reported by Poyatos et al. to find the best-fit line and its equation (Poyatos et al. 2020). Table 3 shows the THC doses, C_max, and the equation and R² value for the best-fit trend line. Using the estimated C_max, 0.35 ng/mL, the K_i determined by Yamaori (1.5 μM), and THC’s molecular weight (314.45 ng/nanomoles), we calculated the ratio as follows:

While our estimated C_max/K_i ratios of CBD and THC were much lower than 0.1, we want to point out that the C_max/K_i ratio of THC in the fourth case report by Brown et al. was 0.012, also less than 0.1, but 100 times closer to the 0.1 ratio. Despite achieving a C_max/K_i ratio of less than 0.1, Brown and colleagues observed interaction between medical cannabis and warfarin (Brown et al. 2021). These conflicting findings confirm the need for additional study of the estimated plasma doses achieved with medical cannabis products and the mechanism by which medical cannabis interacts with other medications.

THC is metabolized by CYP3A4, CYP2C19, and CYP2C9, while CBD is metabolized by CYP3A4, CYP2C19, and potentially by CYP2C9 and CYP1A1/1A2 (Health Canada 2018).Though information about how THC and CBD interact with CYP enzymes varies in the literature, both THC and CBD appear to inhibit CYP3A4, CYP2C19, CYP2C9, and CYP1A1/1A2 (Health Canada 2018; Yamaori et al. 2012). Therefore, it is important that the health care professional monitors for a medical cannabis product’s potential to interact with drugs metabolized by these enzymes, which would include warfarin. Warfarin is comprised of a racemic mixture of S- and R-warfarin, and the S-enantiomer is the more potent of the two. CYP2C9 is responsible for metabolizing the more potent S-warfarin. Therefore, interactions affecting CYP2C9 metabolism are expected to have a greater effect on warfarin’s anticoagulant effects, as measured by INR levels, and the need for warfarin dosage adjustments to maintain therapeutic INR levels. Table 4 summarizes enzymes involved in warfarin, THC, and CBD metabolism.

Numerous variables affect the likelihood of cannabinoids, such as THC and CBD, to interact with CYP450 enzymes, including the route of administration, product formulation, pharmacogenetics, and dosage (Health Canada 2018). Cannabis products may be administered via multiple different routes, including smoking, vaping, and ingestion, and each mode of administration has a unique effect on CYP enzymes and consequently its potential to interact with drugs. For example, smoke from combustion of cannabis contains polyaromatic hydrocarbons which are capable of inducing CYP1A1/2. This could theoretically increase metabolism of R-warfarin leading to decreased INR levels. Sublingual and buccal routes are known to avoid first-pass metabolism by the liver and thus may have less potential for interactions with CYP enzymes. However, in the fourth case report, the authors describe a patient using oromucosal THC products which led to a supratherapeutic INR, suggesting that oromucosal routes do not completely avoid hepatic metabolic pathways (Brown et al. 2021). Despite the use of oromucosal medical cannabis products in our patient, we did not see any changes in INR levels that were suggestive of changes in warfarin metabolism. Therefore, we believe the THC and CBD doses consumed by our patient were below the threshold needed to produce meaningful inhibition of CYP2C9 (Health Canada 2018; MacCallum and Russo 2018; Kaminsky and Zhang 1997).

Finally, we considered the possibility of CYP activity decline as a factor for not observing an interaction between medical cannabis and warfarin. CYP-mediated phase I reactions have been shown to decline with advancing age, but phase II reactions appear to remain intact (Klotz 2009). We are unable to comment on the impact of reduced phase I metabolism on our observed findings. The link between declining CYP-mediated phase I reactions and the risk for medical cannabis-drug interactions may warrant additional investigation.

When evaluating the drug-drug interaction potential of cannabis products, one must take into consideration the route of administration and dosage of THC and/or CBD. Our report was strengthened by the frequent monitoring of the INR level and our ability to verify medical cannabis product information with the patient’s dispensary. Limitations to our clinical assessment include the following: limited information about the patient’s pain as this was primarily managed outside our facility, lack of THC and CBD serum concentrations (test unavailable at our facility, and the information was not needed to effectively manage the patient’s anticoagulation therapy), and the paucity of published data of dose thresholds of CBD and THC expected to interact with warfarin.