As they have stated, hepatic failure is rare if <12 g of paracetamol is ingested.⁴ But the median dose ingested in their study did not reach this level. It is thus unsurprising that the study included only five cases of hepatotoxicity; given this low number, it seems imprudent to claim that no change in incidence has occurred.

In addition, the paracetamol concentrations quoted in the paper are low, considering that concentrations >200 mg/l four hours after ingestion and >130 mg/l six hours after ingestion are used to guide the appropriate use of acetylcysteine treatment. Despite the low concentrations quoted, acetylcysteine was given to 25% and 31% of patients. This does not seem to accord with current practice recommendations.

Robinson et al conclude that the change in pack size had no effect on the incidence of severe liver failure. In our view, a stricter definition of paracetamol overdose is required if we are to make sense of the conflicting reports. Perhaps paracetamol overdose should be confined to subjects who claim to have taken an overdose and have a measurable serum concentration.

We surveyed all admissions for overdose before and after sales of paracetamol were restricted: 116 patients (group A) were admitted in the six months before paracetamol was restricted (February-August 1998) and 112 patients (group B) during February-August 1999. Altogether 52 (45%) paracetamol poisonings occurred before the restrictions were put in place and 40 (36%) after. The number of paracetamol tablets taken was known for 44 patients in group A and 35 in group B. Thirty (68%) patients took more than 16 tablets before rationing and 18 (51%) afterwards; this was reflected in the need for treatment with acetylcysteine, which fell from 16 cases to nine.

Although paracetamol overdose declined, non-paracetamol poisoning increased (from 64 to 72 cases), predominantly involving antidepressant, antipsychotic, and sedative drugs. For both sample periods the average time that each patient spent in hospital was the same (2.6 days). One death occurred from a tricyclic overdose. Although admissions with paracetamol overdose fell, the numbers of cases and bed days needed for care did not alter. Drug selection in self poisoning is usually dictated by what is immediately available, whether obtained over the counter or by medical prescription.^1-5

If paracetamol were a dietary supplement it would have been put into the “most toxic substance on earth” category and banned long ago. Drug interactions with St John's wort (most of which were hypothetical) were acted on with great haste by regulators around the world, despite no deaths ever having been associated with use of St John's wort. Here we have a drug that is available from petrol stations and corner shops and has been killing thousands of people around the world for years, and our regulators turn a blind eye—even when there is a simple solution.