Starting rifampicin in a patient with a mechanical mitral valve on warfarin

Warfarin is given as a racemate. S-warfarin carries most of the anticoagulant effect and is cleared by CYP2C9; R-warfarin is handled by CYP1A2 and CYP3A4.

• Rifampicin is one of the strongest inducers of hepatic CYP3A4, CYP2C9, CYP2C19 and P-glycoprotein. Enzyme induction begins within 2–4 days of the first dose and is maximal at roughly 1–2 weeks, so the interaction is not immediate.
• Induction increases warfarin clearance, so the INR falls. The fall is typically first detectable at day 3–5 and reaches its nadir over the next 1–2 weeks; it persists for the whole 6 months of rifampicin.
• The size of the effect is large: maintenance requirements commonly rise 2–3-fold, and some patients need their dose doubled.
• A complicating factor is isoniazid, which inhibits several CYPs and can push the INR up slightly in the first days. The net early course can therefore be erratic before the induction effect dominates.
• The reverse interaction matters just as much: when rifampicin is stopped at 6 months, induction resolves over 2–3 weeks and the INR climbs steeply on an unchanged dose.

Conclusion: an unchanged warfarin dose will almost certainly leave him subtherapeutic within a fortnight, with a real risk of valve thrombosis and stroke. 'Leaving it alone' is not a safe option.

Start from what cannot be compromised: he has smear-positive cavitary TB, so rifampicin is central to cure and to preventing resistance - it is not the drug to drop.

Options and reasoning:

• Rifabutin substitution is a weaker inducer and would blunt the interaction, but it is reserved for cases where a vitamin K antagonist genuinely cannot be managed; it is still an inducer and is not the first choice for drug-sensitive TB.
• Warfarin with tighter INR control is the preferred route for a mechanical mitral valve. Check INR at baseline, then twice weekly for 2–3 weeks, then weekly until two consecutive values are in range, then no less often than every 2 weeks. Anticipate a 50–100% dose increase, adjusting in 10–20% steps rather than by fixed increments.
• LMWH bridging instead of warfarin for 6 months is possible but is a poor fit: low-molecular-weight heparin is inferior to a VKA for mechanical valves, needs anti-Xa monitoring, and a lorry driver sleeping in a cab cannot reliably self-inject.
• Practical arrangements: register him with a laboratory near his regular routes, allow results to be phoned through, and give him a written dose-adjustment plan to follow only after a result is seen.

Safety netting: written bleeding advice, avoid NSAIDs and aspirin, warn him about macrolides, azoles and the end-of-treatment rise, and give pyridoxine 10 mg daily to prevent isoniazid neuropathy. Monitor LFTs and adherence at each visit.

An INR of 1.6 in a mechanical mitral valve is meaningfully subtherapeutic - the target is 2.5–3.5, and valve thrombosis and systemic embolism are the risks, with stroke the most feared. He is not bleeding, so vitamin K has no role here; giving it would deepen the problem.

Plan:

• Look for a cause before simply raising the dose: missed doses on the road, a new interacting drug, or induction still intensifying.
• Increase the warfarin dose by roughly 10–20% and recheck INR in 2–3 days rather than a week.
• Bridge the gap: therapeutic-dose low-molecular-weight heparin (e.g. enoxaparin 1 mg/kg twice daily) or unfractionated heparin until the INR is ≥2.5 on two consecutive days. Arrange the injections with a practice nurse or pharmacy along his route - unsupervised self-injection is not realistic.
• Advise against the long drive until anticoagulation is therapeutic and he is free of neurological symptoms; any sudden focal weakness, speech disturbance or chest pain means emergency assessment.
• Document the episode and build it into the twice-weekly schedule for the rest of the rifampicin phase, and plan the dose reduction for when rifampicin stops.