Clinical case review
Review the complete Pharmacy case, its five progressive clues, the diagnosis, and the clinical reasoning behind the answer.
Want to test yourself first? Play the original case before reading the clues and explanation below.
WhatsTheDx publishes a new Pharmacy case every day. Five clues, one diagnosis, and no account needed to play.
Play today's Pharmacy caseThese are the case findings in their original reveal order, moving from the broader presentation to the most discriminating evidence.
Clue 1
A 74-year-old woman is brought in by her daughter after three days of nosebleeds that take twenty minutes or more to stop. There are large bruises across both forearms and her left flank that she cannot account for. There was no fall and no injury.
Clue 2
She takes warfarin 4 mg daily for atrial fibrillation, with a target INR of 2.0 to 3.0. She is meticulous about it — she has been in range at every monthly check for four years, and her last reading, a fortnight ago, was 2.4.
Clue 3
Two further things emerge on questioning. Her stools have been black for two days, which she had put down to the iron tablets she is not in fact taking. And eight days ago her general practitioner started a seven-day course of an antibiotic for a urinary tract infection; she finished it yesterday. Nothing else about her medicines, her diet or her drinking has changed.
Clue 4
Her INR is 8.7. Her platelet count is 244 × 10⁹/L — entirely normal — and her haemoglobin has fallen to 96 g/L from a baseline of 131 g/L.
Clue 5
The remaining explanations are closed off one at a time. Liver enzymes, bilirubin and albumin are all normal, so this is not a failing liver. Fibrinogen is 3.4 g/L and the platelet count is untouched, so nothing is consuming her clotting factors. The antibiotic she was given was trimethoprim-sulfamethoxazole.
Diagnosis
The sentence that solves this case is that she has been in range at every monthly check for four years. A patient whose anticoagulation has been stable for that long has not suddenly started taking her tablets wrongly; something was added to the system. Eight days ago that something was a seven-day course of trimethoprim-sulfamethoxazole, and it is one of the named major interactions with warfarin, alongside metronidazole and ciprofloxacin, all of which interfere directly with warfarin metabolism through CYP inhibition. The mechanism is worth holding onto because it explains why the effect is so large. Warfarin works by inhibiting 2,3 vitamin K epoxide reductase, leaving factors II, VII, IX and X inactive along with proteins C and S. It is metabolised by CYP2C9, 1A2 and 3A4, and the S-enantiomer — the one that is 2.7 to 3.8 times more potent — is cleared principally by CYP2C9. Inhibit that pathway and you are not nudging the anticoagulant effect, you are removing the clearance of its most active half. Her INR of 8.7 against 2.4 a fortnight ago is the result, well past the threshold of 3.0 that defines toxicity. The rest of the workup is there to close the doors on everything else that makes an elderly person bleed. A normal platelet count of 244 × 10⁹/L excludes a drug-induced thrombocytopenia. Normal liver enzymes, bilirubin and albumin exclude a liver that has stopped manufacturing clotting factors. A normal fibrinogen alongside normal platelets excludes a consumptive coagulopathy. And the haemoglobin, down to 96 g/L from 131 g/L, with black stools for two days, is what turns this from a number on a screen into an active bleed. That distinction drives the management, and this is where the counterintuitive part sits: the INR itself is a poor predictor of who will actually bleed — in one study of patients with an INR above 10 there was a single major bleeding event, no thrombotic events and no deaths. So the decision is made on the bleeding, not the number. With no significant bleeding and an INR of 10 or below, routine vitamin K is not recommended. She does not qualify for that restraint: she has a falling haemoglobin and melaena, so this is treated as significant haemorrhage with 4-factor prothrombin complex concentrate, intravenous vitamin K 10 mg and consideration of tranexamic acid, accepting the roughly 0.9% thromboembolic rate the concentrate carries. One last point for practice: NSAIDs and SSRIs raise bleeding risk without moving the INR at all, so a reassuring INR never makes them safe company for warfarin.
The explanation and decisive case findings were checked against the source below.
Open the clinical referenceFor education and entertainment only. This fictional case is not medical advice and does not replace supervised clinical training, diagnosis, or treatment.
