Jaundice on TB Treatment and a Smelly Amoxicillin Batch
A 34-year-old man started a standard regimen of rifampicin, isoniazid, pyrazinamide and ethambutol five weeks ago at a TB clinic. He now presents with jaundice, dark urine and nausea. Liver function tests: ALT 486 U/L (upper limit of normal 40), total bilirubin 96 µmol/L, INR 1.4. He denies alcohol but reports drinking a herbal 'liver cleanse' bought from a market vendor for the past two weeks, and has also been taking co-trimoxazole prescribed at another facility for a chest infection. The clinic's pharmaceutical technologist is asked to handle the reporting. Separately that morning, a community pharmacy called to say that three children who took amoxicillin dry powder for suspension from batch AMX-2291 developed vomiting within 30 minutes, and the reconstituted suspension 'smelled sour'.
1. For the jaundice case: is this a reportable adverse drug reaction, how would you assess causality, and what timeline applies?
Read the worked answer
Yes - reportable and serious. Hepatocellular injury with jaundice (ALT well above 3× ULN plus raised bilirubin) meets seriousness criteria and is a recognised safety concern for first-line anti-TB therapy.
Causality assessment using WHO-UMC categories:
- Temporal relationship: plausible. Drug-induced hepatotoxicity typically emerges 2–8 weeks after starting isoniazid, rifampicin and pyrazinamide, and this patient is at week 5.
- Biological plausibility: three of the four drugs are hepatotoxic, and rifampicin is a potent enzyme inducer that alters the metabolism of co-administered drugs.
- Confounders: the herbal preparation (composition unknown, possibly hepatotoxic), co-trimoxazole (cholestatic and hepatocellular injury, particularly at high doses), and the need to exclude viral hepatitis (HBsAg, anti-HAV IgM, anti-HEV) and alcohol.
- Because clear alternative explanations exist alongside the suspect drugs, the most defensible initial category is possible, not probable. Update on dechallenge and outcome.
What to report: the four anti-TB drugs together, plus co-trimoxazole and the herbal product, with the LFT values, the timeline, concomitant medicines, and the dechallenge outcome once the regimen is stopped or modified.
Timeline: expedited reporting - fatal or life-threatening cases within 7 calendar days, other serious cases within 15 calendar days of the information coming to hand. Check the current national pharmacovigilance guidance for the exact intervals and for parallel notification to the TB programme, and inform the facility that prescribed the co-trimoxazole.
2. For the amoxicillin reports: how do you classify the problem, what do you do in the next 24 hours, and how does this pathway differ from the ADR pathway?
Read the worked answer
Classification: this is a suspected quality defect, not primarily a new adverse reaction. Three children vomiting within 30 minutes of a dose, combined with a sour smell from the reconstituted suspension, points to degradation, contamination or a packaging failure. An ADR to amoxicillin is usually rash, diarrhoea or, rarely, anaphylaxis - not a cluster tied to one batch with an odour change.
Next 24 hours:
- Instruct the pharmacy to stop dispensing batch AMX-2291 and to quarantine all remaining stock of that batch, in a labelled area.
- Obtain the full details in writing: batch number, expiry, date of reconstitution, water used, storage before and after reconstitution, number of children affected, symptoms and outcome.
- Retrieve retained samples and arrange submission to the national quality control laboratory through the regulator, with the defect report.
- Notify the regulator immediately via the quality defect/recall channel; ask whether a field alert and recall assessment should begin. Because harm has occurred, this is likely to be treated as a higher-class recall than a cosmetic defect.
- Check whether other batches from the same manufacturer or the same distribution chain are affected, and whether other pharmacies hold batch AMX-2291.
Difference from the ADR pathway: the ADR report describes a patient's reaction to a product assumed to be of acceptable quality and is assessed for causality. A defect report describes a failure of the product or its packaging, is investigated by laboratory testing and batch traceability, and can trigger a recall. When both elements are present - harm plus a product signal - file both; they are not alternatives.
3. Could the two events be related, and how would you prioritise and document them?
Read the worked answer
Related? Almost certainly not. Different products, different patients, and different clinical pictures: a hepatocellular drug reaction in an adult on four hepatotoxic drugs versus acute vomiting in children taking a suspect amoxicillin suspension. There is no shared batch, manufacturer or excipient in the information given. The only common thread is the reporter, which is not evidence of a link.
Prioritisation: the amoxicillin cluster is the immediate public-safety issue. Affected product is still on shelves, other children may be exposed, and action within hours limits the harm. The jaundice case is serious but the patient has already been assessed; a complete, accurate report within the 7–15 day window is the correct priority, not speed at the expense of accuracy.
Documentation:
- Separate files, separate reference numbers, cross-referenced only to note that the same reporter submitted both and that they are not considered related.
- Each file should record what was known, what was done, when, by whom, the decisions taken and the rationale for the causality and classification judgements.
- Note explicitly in the jaundice file that the herbal product is an unregistered, uncharacterised exposure and should also be flagged to the inspectorate as a sale-of-unregistered-product issue.
Wider action: ask whether the market vendor's herbal product is being sold widely, and whether other facilities have seen hepatitis cases with the same product - that pattern would be a signal worth a separate review.