OUP user menu

Paracetamol toxicity: epidemiology, prevention and costs to the health‐care system

DOI: http://dx.doi.org/10.1093/qjmed/95.9.609 609-619 First published online: 1 September 2002


Paracetamol has been used as an analgesic and antipyretic for many years, with toxicity first noted in the 1960s. Since then the incidence of poisoning has increased, and paracetamol is now the most common drug in self‐poisoning, with a high rate of morbidity and mortality. The use, abuse and ways of reducing paracetamol toxicity are reviewed, but in view of the potential for harm, serious consideration should be given to changing the legal status of paracetamol, possibly to a prescription‐only medicine.


Paracetamol (acetaminophen) was discovered in Germany at the end of the 19th century, but was not widely used until midway through the 20th. The toxicity of over‐the‐counter (OTC) analgesics was noticed in the 1960s and 1970s, but paracetamol was considered safe at normal dose. There were few, if any, reports of abuse involving paracetamol and the use of paracetamol steadily increased, replacing the more toxic analgesics available at the time (acetanilide and phenacetin).1 Consumption throughout the world has increased. In the Nordic countries, usage increased five‐fold between 1978 and 1988, and in 1994/95 the rates in some developed countries were >20 g/person/year, although in countries such as the UK, the US, Canada, Australia and New Zealand, consumption was <8 g/person/year.2,,3 In the UK, consumption was reported to have increased from 1500 million 500 mg tablets per year in 1967/68 to 4000 million such tablets in 1993/94.2 An estimate of more recent consumption, including prescribed paracetamol and combination tablets and paracetamol purchased without a prescription, is 3500 million 500 mg tablets in 2000 (IMS Health, Sheen unpublished data). Paracetamol did not undergo the stringent toxicity testing prior to its introduction that now occurs during drug development. It was not until 1966 that hepatotoxicity due to paracetamol was first reported in humans.4,,5

Is paracetamol a tolerable and effective drug?

Paracetamol is now used in many forms either alone or in combination with other drugs (usually opiates) for analgesia and in other mixtures such as cold ‘cures' for its analgesic and antipyretic properties. Not only must a drug be effective but it must also be tolerable, i.e. without severe or excessive side‐effects. Only one randomized controlled study has directly compared the tolerability of paracetamol to those of aspirin and ibuprofen.6 This study concluded that paracetamol was no better tolerated than ibuprofen, although it was better tolerated than aspirin. Since 1990, a number of trials have shown paracetamol to be an effective antipyretic, but its analgesic effectiveness compared to other painkillers, including ibuprofen (another frequently used and readily available analgesic) is variable (Tables 1 and 2).1,7–,109 In adults, paracetamol can only be considered a mild to moderate analgesic.

View this table:
Table 1

Summary of results of randomized controlled trials using oral or rectal paracetamol

Acute pain (n=76)Chronic pain (n=17)References
Better than placebo146Acute pain 16–29, chronic pain 30–35
Worse than placebo00
Equivalent to placebo60Acute pain 36–41
Better than another drug30Acute pain 42–44
Worse than another drug2710Acute pain 17,19,23,28,45–67, chronic pain 31,33,34,68–74
Equivalent to another drug256Acute pain 18,20,24,25,27,39,40,75–92, chronic pain 30
No effective analgesia120Acute pain 97–108
  • Totals are greater than n, as some studies compare paracetamol with another drug and placebo.

View this table:
Table 2

Randomized controlled trials comparing ibuprofen with paracetamol

AuthorDateIbuprofen doseParacetamol doseResultPain typeNotesReference
Behotas S1992400 mg1000 mgIbuprofen betterAcuteSingle dose47
Bradley JD19911200 or 2400 mg4000 mgEquivalent effectChronic93
Bradley JD19921200 or 2400 mg4000 mgEquivalent effectChronic94
Bradley JD20011200 or 2400 mg4000 mgEquivalent effectChronicFurther analysis of Bradley 199296
Milgrom C19932400 mg3000mgNo effectAcute103
Packman B2000400 mg1000 mgIbuprofen betterChronicSingle dose using soluble ibuprofen33
Schachtel BP1996400 mg1000 mgIbuprofen betterChronicSingle dose using soluble ibuprofen74
Torabinejad M1994500 mg500 mgNo better than placeboAcute41

How common is paracetamol poisoning?

Hospital admissions due to poisoning have steadily increased from the 1950s.110 Since the mid‐1970s. there has been an increase in the number of paracetamol overdoses, such that paracetamol has now become the substance most frequently used in deliberate self‐poisoning in the UK.111 In Oxford, UK, the proportion of overdoses with paracetamol increased from 14.3% in 1976 to 42% in 1990, and in 1993, 47.8% of all overdoses involved paracetamol or paracetamol‐containing drugs.111,,112 It has also become increasingly common in other countries including Denmark and Australia.113,,114 In Scotland, the rate of paracetamol overdose increased almost 400% between 1981–83 and 1991–93, and was higher in more deprived areas.115,,116

Paracetamol overdose is a significant cause of hospital admission, but severe liver damage is infrequent, and when it does occur the prognosis is generally good. A study in one hospital in Western Australia between 1985 and 1990 found 306 admissions for paracetamol overdose. Severe liver injury (defined as an alanine transaminase >45 IU/l, a prothrombin time >18 s and encephalopathy) occurred in 6.9%, but all recovered with supportive therapy and no patient required a liver transplant.117 However, in a study of patients admitted to a specialist liver failure unit in the UK between 1987 and 1993 with a diagnosis of acute liver failure due to paracetamol poisoning, only 30% fulfilled clinically defined transplant criteria.118 Patients with poorer outcomes and those listed for transplant tended to present later to hospital and to have taken a larger overdose.114,,118 In a prospective trial of 80 patients admitted between 1992 and 1993, 25 had acute liver dysfunction (as defined by an INR >1.2 with or without abnormal liver function tests) following consumption of more than 25 tablets (12.5 g) of paracetamol. Of the 80 patients, 60% had obtained tablets from blister packs, 46% from loose preparations and 6% of the patients had used both types.111

In 1977, 144 deaths in England and Wales were attributed to paracetamol.110 Between 1993 and 1997, this had increased to approximately 500 deaths attributable to drugs containing paracetamol per year in England and Wales.119 Over the same period of time, the mean annual number of deaths from road transport accidents (RTA) was 3291 (data from Office for National Statistics; www.statistics.gov.uk). There is much public concern and Government action to reduce RTA deaths, but little publicity is given to a minor analgesic that caused a sixth as many potentially avoidable deaths as RTAs.

While some drugs have limitations imposed on their availability for small adverse risks, paracetamol is not looked upon with the same critical concern. An example of this is terfenadine. In 1997, its availability was changed from OTC to prescription‐only, due to the risk of cardiac arrhythmias. However, an observational historical cohort study calculated that more cardiac arrests and ventricular arrhythmic events occurred with other OTC antihistamines compared to terfenadine. Also, there was no increase in the risk of life‐threatening ventricular arrhythmias when terfenadine was compared to ibuprofen.120

Accidental or intentional?

Although most paracetamol poisoning is due to the wilful ingestion of supratherapeutic doses with the intention of committing suicide, a proportion of patients present with toxicity due to incorrect dosing for therapeutic intent or accidental overdose. In children under 5 years old, ingestion of toxic substances is almost invariably unintentional, and is usually due to household chemicals or drugs belonging to other family members.121 In a retrospective study in America, reviewing medical charts for 10 years between 1988 and 1997 of children (aged <18 years) presenting with paracetamol overdose, 53% presented with unintentional ingestion and 3% with toxicity due to a dosage error. The unintentional ingestion was seen in young children, from 2 months to 13 years, whereas intentional overdose was seen in the teenagers, the majority of whom were female.122 In another American study retrospectively analysing hospitalizations for excessive paracetamol ingestion, 30% were considered accidental. The dose taken was half that of those attempting suicide, but over half of the subjects presented more than 24 hours after ingestion (compared to only 14% of the suicidal patients). Those with accidental poisoning had a higher incidence of hepatic coma and death, and also spent longer in hospital.123 In the UK. only 8% of patients with acute liver failure stated they had accidentally taken an overdose, and most of these patients presented to hospital >24 h after the event.118 The difference between the UK and USA is difficult to explain, but may be due to cultural responses to overdose, healthcare funding differences between the two countries and consequent reporting pressures of the two systems.

Why is paracetamol used in attempted self‐harm?

Paracetamol is the most frequently used drug by those who take an intentional overdose in the UK.111 However, only 0.1% of all paracetamol overdoses resulted in death in the US, thus paracetamol can hardly be considered to be an effective drug with which to commit suicide.124 Of course, paracetamol can be readily purchased on the high street without the need for a prescription. In Sweden, as sales of analgesics increased so did the suicide rate (although this relationship was not shown with paracetamol alone in this study, possibly reflecting successful treatment of paracetamol overdose).125 Another study showed a relationship with the sale of paracetamol and the numbers of paracetamol overdoses in France and the UK.126

Since paracetamol has been available OTC, the public's understanding of the effects of a paracetamol overdose have changed. In 1976, patients were not aware that there was a delay of several days before the onset of serious symptoms, and Gazzard et al. found that had they known this, they would not have used it.127 A second study in the early 1990s found that most patients recognized that paracetamol could cause death, but again, most were unaware of the delay in onset of severe symptoms and many thought that overdose would cause unconsciousness. This study also confirmed that people took paracetamol because it was easily available and inexpensive.128 School children between the ages of 12 and 19 in the UK and the US were aware that paracetamol could be harmful or fatal although the dose required was greatly overestimated by over half the children. Belief that sedation was a side‐effect was also common.129 It seems that education has informed the population about the risk of death with paracetamol overdose, but not the unpleasant way in which it comes about.

Advertising and awareness of the drug may influence its use. Newspaper reports and television programmes about a suicide may cause a transient rise in attempted suicide.130,,131 Equally, the media can be a major source of education.132

Measures to reduce toxicity

In one study, patients admitted with a paracetamol overdose were asked about factors that might have deterred them from taking the overdose. Although 66% would still have used paracetamol with the knowledge that it could cause death, only 35% would still have used it had they known that the harmful effects could be delayed for several days. Warning labels (‘paracetamol can cause death’) would only have deterred 25%. Repackaging into blister packs would have had little effect, and limiting the number of tablets would only have led to 37% taking fewer tablets or not overdosing. If paracetamol was made prescription‐only, 35% would not have taken an overdose, and 40% would have sought an alternative. If an antidote were contained within the paracetamol tablet, 64% would not have taken the overdose.111 Paracetamol use in the UK and France was compared between 1974 and 1990 to assess whether restrictions were justified in the availability of the drug.126 At the time, paracetamol was freely available in England and Wales in unlimited quantities from a pharmacy, and up to 12 g from supermarkets. In France, paracetamol could only be purchased from pharmacies and the contents of each pack was legally limited to 8 g. There was a strong correlation in the UK and France between paracetamol sales and paracetamol overdose. Case fatality rates, however, were four times higher in England and Wales than in France. They concluded that the greater availability of paracetamol was paralleled by increases in its use with non‐fatal and fatal overdose, and suggested that pack size restrictions should be stricter in the UK. In September 1998, the Medicines Control Agency restricted the sale of OTC paracetamol in the UK. Pharmacies can sell packs containing a maximum of 32 tablets (16 g) although up to 100 tablets may be sold at the discretion of the pharmacist. In outlets other than pharmacies, packs with up to 16 tablets (8 g) can be sold.133

Effectiveness of the UK pack size restriction

Recent research in England and Wales has suggested that there has been a reduction in the number of patients listed for liver transplant and of overdose since September 1998.134–,136 However, one study looked only at severe overdoses referred to a specialist liver unit, and did not consider less severe overdoses, while another reported only a small number of cases and so may not accurately represent outcome.134,,137 In Tayside, Scotland, using serum paracetamol levels as a marker of true toxicity, there was no effect from the pack size reduction.138 There has also been no change in the number of patients being referred to the transplant unit in Scotland.139 Another study also commented that there was no change in the mean highest serum paracetamol concentration before and after the pack size reduction, although this is not commented upon by the authors in the final conclusion.136

Methionine‐paracetamol combination tablets

Methionine is a glutathione donor required in the metabolism of paracetamol. Combination tablets were first suggested in 1974, and released on the UK market in 1987.140,,141 The analgesic effectiveness of paracetamol has been shown in the combination drug, with a slight increase in minor adverse effects such as drowsiness,142 Other adverse effects of methionine are mild, and include nausea, flatulence and headache.143 As most paracetamol overdoses are impulsive methionine would have to be added to all formulations containing paracetamol. This would mean that a lot of people would needlessly consume methionine, and the dose required to be effective is unclear. In rats, who are quite resistant to the hepatotoxic effects of paracetamol, a methionine dose of 10% of the paracetamol dose was protective,144 but an equivalent effective dose cannot ethically be determined experimentally in humans. In the UK the only paracetamol‐methionine combination available is Paradote (Penn Pharmaceuticals). This contains 100 mg methionine and 500 mg paracetamol (i.e. 20% methionine). Concern has been expressed that long‐term methionine intake could promote carcinogenesis.145 Methionine is metabolized to homocysteine, and raised plasma homocysteine concentrations have been associated with endothelial dysfunction,146 stroke,147 and coronary heart disease.148 The dose at which these effects occur is not clear. In healthy adults, 100 mg/kg L‐methionine caused a significant increase in homocysteine levels.149,,150 In a study looking at the effects of dietary methionine and homocysteine levels in healthy males, an intake of a mean of 2112 mg of methionine a day for 7 days had no significant effect on plasma homocysteine levels.151 Thus the dose needed to have an effect on homocysteine levels would appear to be considerably greater than the 800 mg that would be ingested with the 4 g recommended maximum daily dose of paracetamol.

Another paracetamol‐methionine combination tablet was withdrawn from the UK market after it was placed on the list of drugs not available for prescribing at NHS expense. It was felt that it was expensive, there were alternative drugs to paracetamol available and that allowing the product to be prescribed would have no effect on self‐poisoning if the paracetamol tablets were obtained OTC.

The net cost per tablet of Paradote is over six times that of generic paracetamol. This would perhaps deter a person planning an overdose, if it were the only form of paracetamol available. The cost might also limit the availability of an effective analgesic for appropriate use to much of the population.

There do seem to be risks of using methionine which need to be evaluated further. Until these risks are elucidated, it would seem ethically unacceptable to require all people to take a paracetamol‐methionine combination. Cost and government policy will also affect the availability and use of these types of drugs.

Alternative drugs

In adults there is the possibility of substituting an alternative drug such as a non‐steroidal anti‐inflammatory drug (NSAID). These have been shown to be at least as effective as paracetamol. There are however, concerns over the adverse effects of these drugs including, in particular, gastrointestinal bleeding. Over the last few decades there have been many studies of the risks of haemorrhage. A recent review of these has highlighted that the risk of bleeding can be stratified (age, sex and history of previous peptic ulcer) and by the type of NSAID.152 This study did not include the selective cyclo‐oxygenase 2 inhibitors (e.g. Rofecoxib and Celecoxib). Recent studies of these drugs has shown that they are effective analgesics, with a lower incidence of peptic ulcer bleeding.153–,156 These newer drugs may provide a safe and effective alternative to paracetamol, although recent concern over their cardiovascular safety profile has rather soured their attraction.157,,158

Cost to healthcare systems

The cost of poisonings and the subsequent treatment is significant. In 1985, the lifetime direct and indirect costs associated with all poisonings in the US was estimated at $8.4 billion.159 In the US, the direct cost of paracetamol overdose has been estimated at $87 million annually, and this is likely to be a conservative estimate. In 1995, it was calculated that the average cost associated with intentional paracetamol poisoning involving adolescents and adults in the US was $2172 per case. This allowed for emergency department visits and treatment with N‐acetylcysteine, and also averaged out the additional costs from hepatic injury, hepatic failure and the need for liver transplantation, but again is probably a conservative estimate.160 However, in a children's hospital in Boston, USA the cost per case of paracetamol poisoning (excluding physician's fees) ranged from $17 349 in 1992 to $7080 in 1995, the reduction in costs being due to reduced in‐patient length of stay.159 In a study looking at admissions between 1992 and 1995, the mean cost of care in patients having taken an accidental overdose was $19 000, compared to $8500 in patients attempting suicide.123 This difference is likely to represent the longer length of stay and the cost of treating complications such as hepatic failure. There are no similar data for the UK, as health care is not charged for in the same way as the US. However, it seems likely that the cost burden will be as significant in the UK as in the US and this is supported by a small study in Tayside, although the overall costs in Scotland are lower than in the US.161 The mean direct cost per case in Tayside was £181 (excluding costs of specialist care of acute hepatic failure and liver transplant). If this cost were to used to represent the UK, then the annual direct cost of hospitalization with paracetamol poisoning would be approximately £8 million.


Paracetamol is a commonly used, moderately effective analgesic and antipyretic. In overdose it causes significant morbidity and mortality. The burden to health care services is considerable, with a high financial cost and many hospital admissions. According to the Medicines Control Agency Medicines Act Leaflet (MAL 82, March 1996), which gives guidance on changing the legal classification of a medicine to the General Sale List, a criterion for inclusion on the General Sale List is: ‘where the hazard to health, the risk of misuse, … is small and where wider sale would be a convenience to the purchaser’ (our italics). It is surprising that paracetamol is available on the General Sale List, as it appears to fail this criterion for an OTC medication.

Present approaches to reduce toxicity have had variable and moderate effects. Other methods of reducing this burden must be considered. These could include public education on the effects of overdose, further research on drug‐antidote combination tablets and changing the legal status of adult doses from the General Sales List to the prescription‐only medicines list, or at least restricting it to pharmacy‐only sales.


  • Address correspondence to Dr C.L. Sheen, Department of Clinical Pharmacology, Ninewells Hospital and Medical School, Dundee DD1 9SY. e‐mail: chris{at}memo.dundee.ac.uk


View Abstract