Warfarin and other oral anticoagulants

This chapter is relevant to Section Q2(i) of the 2017 CICM Primary Syllabus, which expects the exam candidates to "understanding of the pharmacology of anti-coagulants, anti-platelet drugs, thrombolytic drugs and anti-fibrinolytic drugs". It is specifically relevant to three SAQs:

• Question 10 from the first paper of 2021
• Question 10 from the first paper of 2016
• Question 8 from the first paper of 2015

None of these questions were especially inventive in their cruelty, making this a fairly straightforward exercise in pharmacology. Additionally, Question 21 from the first paper of 2014  asked the trainees to compare dabigatran and warfarin; and that seems like an excellent excuse to drag these substances under the light. They were also used in the Fellowship exam, also in the first paper of 2015 (Question 4), except the Second Part candidates were being asked about the various advantages and disadvantages of their use in atrial fibrillation.

For a bit of contrast, specific sources for this information that have some kind of scientific credibility include:

• Tran et al (2014) is an excellent overview of oral anticoagulants
• Ansell et al (2004) or  Ufer (2005) give a brief tour of warfarin and its chemical relatives
• AHA/ACC/HRS Guidelines for AF (2014) - mainly to help answer Question 4 from the first paper of 2015 with some specific recommendations.
• Stangier et al (2009) - pharmacology of dabigatran
• Kreutz (2010) - pharmacology of rivaroxaban
• Frost et al (2013) - pharmacology of apixaban

Chemical properties and chemical relatives of warfarin

Warfarin is a derivative of coumarin, presented as a racemate of two active enantiomers, of which the S-enantiomer is the most active. Coumarin is a benzopyrone commonly found in all sorts of plants, and it has no anticoagulant effects on its own. Warfarin itself is a derivative of a derivative, originating from dicoumarol (the clover-derived toxin responsible for millions of cattle deaths in the 1920s). There are multiple other coumarin-based rodenticides and anticoagulants (the two categories overlap to a surprising degree), of which the most familiar names will probably be acenocoumarol, ethyl biscoumacetate, phenprocoumon and the newer more lethal long-acting versions like bromadiolone and brodifacoum.  All of these drugs are related both by their chemical structure and by their mechanism of action, which is often referred to (inaccurately) as "vitamin K antagonism".

Non-warfarin oral anticoagulants

To call these drugs "NOACs" would be to show one's age, as they are no longer Novel by any means. "DOACs", or "Direct oral anticoagulants", might seem more accurate, as their mechanism is more direct than warfarin (i.e. they usually directly interact with the clotting cascade instead of trying to undermine it by sabotaging the synthesis of critical factors. Chemically, they are a rather diverse group. Dabigatran is a small molecule polypeptide similar to argatroban, whereas apixaban and rivaroxaban are oxazolidinones, chemically related to linezolid.

Pharmacokinetics of oral anticoagulants

For such a chemically diverse group of agents, they are remarkably similar in their pharmacokinetics. The reasons for this are an accident of marketing, and they make all of their properties highly predictable. All of these oral agents would have to be well absorbed orally and to have a decent oral bioavailability to be sold as tablets, which means they would all have to be highly lipid soluble (and poorly water soluble), and this lipophilicity would make them all highly protein bound. 

Name	Absorption and bioavailability
Warfarin	Rapidly absorbed High oral bioavailablility (~ 100%) Maximum blood concentration about 90 minutes after administration
Dabigatran	Available as dabigatran etexilate, which is a pro-drug; dabigatran on its own is mch too polar to be absorbed effectively. The exetilate is rapidly absorbed High oral bioavailablility (~ 100%) Maximum blood concentration about 90-180 minutes after administration
Rivaroxaban	Rapidly absorbed High oral bioavailability (~ 80-100%) Maximum blood concentration about 120-240 minutes after administration
Apixaban	Rapidly absorbed Oral bioavailability ~50% Maximum blood concentration about 90-200 minutes after administration

Absorption and bioavailability for most of these drugs is not a problem. Warfarin is rapidly and completely absorbed in the intestine, reaching its maximum concentration rather quickly. Unfortunately, because of its pharmacodynamics, this does not really translate into any sort of speed advantage when it comes to the onset of its effect. The only exception is dabigatran: itself, the drug  would not be absorbed at all unless it was presented as an etexilate: otherwise, it would be fully ionised at gastric and intestinal pH, making it very difficult for it to cross into the bloodstream. The etexilate however gets absorbed effortlessly, giving dabigatran an effective bioavailability of basically 100%.

 pKa for all of the oral anticoagulants is unfavourable to good water solubility. None of them are especially soluble in water, and none have a conveniently available mechanism of parenteral delivery. 

Name	pKa and solubility	Volume of distribution and protein binding
Warfarin	pKa 5.87; practically insoluble in water	VOD=0.08-0.12, 99% protein-bound
Dabigatran	pKa 4.0, solubility in water is not very good	VOD=1.0L/kg; 35% protein-bound
Rivaroxaban	pKa 13; practically insoluble in water	VOD=0.7L/kg; 95% protein-bound
Apixaban	pKa 13.1; practically insoluble in water	VOD=0.3L/kg; 87% protein-bound

Distribution and protein binding:  All oral anticoagulants have a fairly small volume of distribution, as most of them are highly protein-bound to plasma proteins. 

Metabolism and elimination for warfarin and rivaroxaban is mainly hepatic, resulting in inactive metabolites which are really excreted. In contrast, dabigatran and apixaban rely considerably (for dabigatran, entirely) on renal clearance.

Name	Metabolism	Elimination	Half life
Warfarin	The more potent S-isomer is metabolised by hepatic CYP2C9 to 7-hydroxywarfarin, an inactive hydroxylated metabolite. Another route is through reductases, into reduced metabolites (warfarin alcohols) with minimal anticoagulant activity.	All of the metabolites ultimately make their way out in the urine. 92% of the radiolabelled dose is recovered in urine after 1 week. Minimal free warfarin is renally excreted	Terminal half-life of warfarin after a single dose is approximately 1 week; Effective half-life ranges from 20 to 60 hours Mean half-life is about 36-42 hours.
Dabigatran	Mainly renally excreted as unchanged drug; but about 20% is conjugated with glucuronic acid to form acylglucuronides. These conjugates are pharmacologically active and demonstrate almost identical properties of free, unconjugated dabigatran.	80% renal excretion of unchanged drug 20% biliary excretion of acylglucouronides	Half life = 12-14 hours
Rivaroxaban	66% is metabolised in the liver: a substrate of CYP3A4 and CYP2J2. Inactive metabolites are excreted by biliary (50%) and renal (50%) routes	33% renal excretion of unchanged drug 66% hepatic metabolism	Half life = 7-12 hrs
Apixaban	33% is metabolised in the liver: a substrate of CYP3A4/5 Inactive sulfate conjugates are renally excreted	Hepatic metabolism, renal excretion and biliary secretion are each responsible for elimination of approximately one-third of dose. 33% renal excretion of unchanged drug 33% biliary and intestinal secretion of unchanged drug 33% hepatic metabolism	Half life = 12 hours

Mechanism of the anticoagulant effect

"Vitamin K inhibitors" are so named because it would probably be inconvenient to call them "vitamin K epoxide reductase complex inhibitors". They sure don't inhibit vitamin K in any meaningful sense. Vitamin K is an essential cofactor for the activation of factors II, VII, IX, and X. These factors are synthesized as precursors, and require post-translational carboxylation by gamma glutamyl carboxylase. Reduced vitamin K is required as a cofactor for this reaction. In its absence, the secreted factors are inactive, and the extrinisic pathway is affected.

The lack of factor 7 does not disturb the intrinsic pathway overmuch; or at least when it is tested using various clay-based surrogates, it is still possible to have a relatively normal clotting time. As the result, PT (and its normalised version, the INR) are used to monitor the effects of warfarin. Still, APTT can be affected, as it tests the function of factors II, X and IX. Kearon et al (1998) found that on average it tends to increase by about 16 seconds for every 1.0 increment in INR.

Direct thrombin inhibitors interact with the active site of thrombin and inactivate it. The result is that thrombin, including fibrin-bound thrombin, is not able to do any of its normal duties. We may recall that those duties are central to the clotting cascade. Thrombin is the final common step for both the intrinsic and the extrinsic pathway, as it is responsible for multiple essential steps:

• It mediates the conversion of fibrinogen to fibrin.
• It is also a potent stimulant of platelet activation
• It enhances clot stability by activating the intrinsic pathway
• It cleaves Factor VIII off from Von Willebrands Factor  

Dabigatran and argatroban attack this central factor, disabling the clotting cascade. PT and APTT do not work particularly well to monitor the effect, as they tend to remain normal while the patient is clinically fully anticoagulated. Di Nisio et al (2005) propose ecarin time as a better option.

Factor Xa inhibitors attack the pre-thrombin step in the clotting cascade. Factor Xa is the final form of Factor X, which is activated by either the extrinsic or intrinsic pathway in the course of the "amplification" stage of haemostasis.  Factor Xa then joins Factor Va to form the prothrombinase complex, which converts prothrombin into thrombin and the rest is a well-known cascade of events that leads to the formation of fibrin (Yeh et al, 2012, probably explain it better). In effect, to disable the function of prothrombinase would be the same as to disable thrombin itself. 

Reversal agents

Reversal of warfarin

Time. Literally waiting for some halflives to wear down may be enough. Patients with an elevated INR and no actual bleeding can be managed conservatively.  The problem with giving any dose of vitamin K is that eventually, you will probably need to rewarfarinise the patient, and it will be more difficult to titrate their dose if their bloodstream ends up being full of active factors, so waiting is often a reasonable option. However, that is too boring for the CICM examiners and they typically ask for the pharmacological measures. In other words, read the question carefully: this safe and reasonable approach may not score any marks.

Vitamin K (phytomenadione) is the reversal agent for warfarin. It increases the availability of the reduced substrate for the activation of clotting factors, which means you are not as reliant on the blocked epoxide reductase. Do not be fooled, the reductase remains blocked, but at least now you have some factors getting activated, which means normal clotting function will be possible as soon as the liver produces enough of them. That is the main caveat: this is not an instant solution. It is more rapid than just waiting for the warfarin to wear off, but the acutely bleeding patient may require something more.

Additionally, the dose matters. A non-urgent situation may call for a smaller dose, potentially delivered as a tablet. Oral Vitamin K has enough oral bioavailability to be effective in this setting. The college examiners laboured the point that this fat-soluble vitamin will require bile salts to absorb, which excludes some people from being able to benefit (eg. those who have undergone a cholecystectomy, or those with end-stage liver disease). A larger dose is called for if the patient is at considerable risk of bleeding to death. The IV formulation may be called in if time is of the essence or if the patient is unable to absorb it enterally; in which case it is still usually well tolerated apart from occasional case reports of anaphylaxis to the benzyl alcohol excipient (Aziz et al, 1996) 

Prothrombinex contains factors II, IX, X and low levels of factor VII. The dose is 25 – 50 IU/kg. This blood product concentrate rapidly restores the circulating levels of the necessary factors, and with a lot less infused volume than the next best option, which is...

 FFP of which the maximum dose is 15ml/kg. Consider that this could be 1000ml of a highly hyperoncotic solution, funnelled into the patient rapidly to reverse their coagulopathy. The outcome may be TACO, transfusion-associated circulatory overload. For this reason, and damn the expense, prothrombinex is a much better option in virtually every case.