Drug Safety 2005; 28 (3): 227-240

REVIEW ARTICLE 0114-5916/05/0003-0227/$34.95/0

 2005 Adis Data Information BV. All rights reserved.

Tolerability of Paracetamol
Garry G. Graham,1 Kieran F. Scott2 and Richard O. Day1
1 Department of Clinical Pharmacology, St Vincent’s Hospital and Department of Physiology
and Pharmacology, School of Medical Sciences, Sydney, Australia
2 St Vincent’s Hospital Clinical School, University of New South Wales, Sydney, Australia

Contents
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
1. Identification of Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
2. Metabolism of Paracetamol (Acetaminophen) Relevant to its Toxicity . . . . . . . . . . . . . . . . . . . . . . . 228
3. Hepatotoxicity of Paracetamol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
3.1 Hepatotoxicity from Overdosage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
3.2 Hepatotoxicity from Therapeutic Doses in Adults? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
3.3 Hepatotoxicity from Therapeutic Doses in Children? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
3.4 Hepatotoxicity from Therapeutic Doses in Patients Drinking Alcohol? . . . . . . . . . . . . . . . . . . . . . . 233
3.5 Use of Paracetamol in Alcoholics and in Patients with Liver Diseases . . . . . . . . . . . . . . . . . . . . . . 234
4. Labelling and Packaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
5. Gastrointestinal Tolerability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
6. Renal Tolerability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
7. Haemostasis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
8. Hypersensitivity Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
9. Occurrence of Cancers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
10. Tolerability During Pregnancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
11. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238

Abstract The excellent tolerability of therapeutic doses of paracetamol (acetaminophen)

is a major factor in the very wide use of the drug. The major problem in the use of
paracetamol is its hepatotoxicity after an overdose. Hepatotoxicity has also been
reported after therapeutic doses, but critical analysis indicates that most patients
with alleged toxicity from therapeutic doses have taken overdoses. Importantly,
prospective studies indicate that therapeutic doses of paracetamol are an unlikely
cause of hepatotoxicity in patients who ingest moderate to large amounts of
alcohol. Controlled clinical trials have found that paracetamol is very well
tolerated by the gastrointestinal tract. While variable results have been found in
case control studies, most studies have shown no change or a small increase in the
relative risk of perforations, ulcer or bleeding in the upper gastrointestinal tract.
However, associations between the use of paracetamol and gastrointestinal toxici-
ty, as well as with chronic renal disease and asthma, are very likely to reflect
biases in some case control studies. In particular, such biases may be caused by the
perceived high tolerability of paracetamol in these diseases. The consequent use
of paracetamol in these diseases states then leads to an apparent association
between paracetamol and the disease. Despite metabolism of paracetamol to
reactive compounds, hypersensitivity reactions are rare, although urticaria occurs
in occasional patients. Paracetamol appears to be well tolerated during pregnancy
although prospective studies are required.
228 Graham et al.

Paracetamol is a very well tolerated drug at thera- poses a problem in the evaluation of adverse reac-
peutic doses (i.e. up to 4 g/day in adults) and this tions to analgesics and is a particular problem with
excellent tolerability is a major factor in its wide paracetamol because of its widespread use.[1]
usage. The aim of this review is to survey the Adverse reactions to NSAIDs are largely due to
adverse reactions to therapeutic dosages of paraceta- their inhibition of the synthesis of prostaglandins.
mol. In particular, we have examined the conclu- The non-selective NSAIDs, such as aspirin (acetyl-
sions of recent reviews on these reactions and com- salicylic acid) and indometacin, inhibit prostag-
mented on the clinical significance and actual occur- landin synthesis by blocking both cyclo-oxygenase
rence of these adverse reactions. (COX)-1 and COX-2. The gastrointestinal damage
Hepatotoxicity after overdosage of paracetamol produced by these drugs is largely attributed to their
is well known and is the major problem with use of inhibition of the production of cytoprotective pros-
this drug. Conversely, hepatotoxicity after therapeu- taglandins in the gastrointestinal tract.[2] This is pri-
tic dosages appears to be highly unusual. Although marily a COX-1 dependent pathway and the COX-2
many case histories have been interpreted as indicat- selective drugs are better tolerated by the gastroin-
ing that therapeutic doses of paracetamol can cause testinal tract than are the older non-selective
hepatoxicity, recent critical reviews indicate that drugs.[2] In intact cells stimulated by cytokines,
hepatotoxicity from therapeutic dosages of paraceta- paracetamol inhibits prostaglandin synthesis when
mol is much less common than is widely believed. the latter is probably mediated by COX-2.[3] Thus,
Controlled prospective studies generally indicate the not surprisingly, the gastrointestinal and overall tol-
safety of therapeutic doses of paracetamol to the erability profiles of paracetamol, apart from hepato-
liver as well as to the gastrointestinal tract, although toxicity, are similar to those of the selective COX-2
mild toxicity may occur in the respiratory tract. By inhibitors.
contrast, adverse reactions in the gastrointestinal
tract have been indicated by epidemiological stud- 1. Identification of Studies
ies. Examination of these discordant results and
conclusions is a significant aspect of the present Relevant studies and review papers for examina-
review. tion were mostly identified from the Medline
database (1975– January 2005), using the keywords
Analysis of epidemiological studies of paraceta-
‘acetaminophen’ or ‘paracetamol’, together with
mol indicates that several studies have biases that
‘adverse effects’, ‘side effects’, ‘tolerability’, ‘gas-
may very well invalidate conclusions drawn from
trointestinal’, ‘stomach’, ‘hepatic’, ‘liver’, ‘gluta-
them about the incidence of adverse reactions. One
thione’, ‘metabolism’, ‘renal’, ‘kidney’, ‘sodium’,
bias is confounding by indication. In the case of
‘pregnancy’, ‘asthma’, ‘hypersensitivity’, ‘pros-
paracetamol, this confounding may occur when
taglandin’, ‘cancer’ or ‘dose’. The reference lists of
paracetamol is selected by prescribers or by pur-
recent reviews (2000–January 2005) and research
chasers in preference to NSAIDs because of the
papers were also used to identify relevant papers.
perceived greater safety of paracetamol with respect
to a particular organ system (e.g. the gastrointestinal 2. Metabolism of Paracetamol
tract).[1] This type of bias, although well recognised, (Acetaminophen) Relevant to
is difficult to eliminate when full patient histories its Toxicity
are not known to the investigators, as is often the
case. The bulk of the metabolism of paracetamol is by
A second problem is protopathic bias.[1] With formation of the glucuronic acid and sulfate conju-
respect to adverse reactions to drugs, this term is gates (figure 1). These two conjugates account for
used to describe the occurrence of disease that starts about 80% of the elimination of paracetamol. In
to develop before the administration of the drug. The addition, a small amount is excreted unchanged.
drug is then started by the patient in an early phase None of these processes is associated with toxicity,
of the disease that continues to develop and is then apart from occasional thrombocytopenia possibly
associated with use of the drug. Protopathic bias because of paracetamol sulfate. By contrast, it is

 2005 Adis Data Information BV. All rights reserved. Drug Safety 2005; 28 (3)
Tolerability of Paracetamol 229

NHCOCH3 NHCOCH3
NHCOCH3

−
OG OSO3
OH
Glucuronide Sulfate
Paracetamol

CYP2E1
COX/MPO CYP2A6 COX/MPO
+ GSH NCOCH3

NHCOCH3
Semiquinone imine
free radical intermediate

O
OH
NHCOCH3 NHCOCH3 NAPQI
OH
3-hydroxy
paracetamol NHCOCH3

OH OH
NHCOCH3
Paracetamol dimer
SG
OH
Glutathione
OCH3 conjugate

OH
3-methoxy
Cysteine and N-acetyl
paracetamol
cysteine conjugates

Fig. 1. Pathways of the metabolism of paracetamol (acetaminophen). The major metabolites involve non-oxidative metabolism to the
glucuronide (G = glucuronyl) and sulfate conjugates that account for approximately 50% and 30% of doses of paracetamol, respectively.
Cytochrome P450 (CYP) 2E1 oxidises paracetamol by a two electron oxidation process to N-acetyl-p-benzoquinone imine (NAPQI), which
is subsequently hydrolyzed to the cysteine adduct and acetylated. CYP2A6 oxidises paracetamol to 3-hydroxy paracetamol, which is
methylated. Myeloperoxidase (MPO) and the peroxidase functions of cyclo-oxygenase (COX) isoenzymes convert paracetamol by overall
one electron and two electron oxidation processes. The one electron oxidation product is the semiquinone free radical intermediate that is
either reduced back to paracetamol by reduced glutathione (GSH) with the simultaneous oxidation of GSH, or converted to paracetamol
dimer and small amounts of higher polymers. The dimer and further polymeric products have been detected in vitro but have not been
examined in vivo. NAPQI is also formed by MPO and COX isoenzymes.

well recognised that oxidative metabolism of parac- conjugate and then to N-acetyl conjugate, which are
etamol can lead to hepatotoxicity.[2] both excreted in urine (figure 1).[2]
The cytochrome P450 (CYP) system in the liver CYP2E1 is the CYP isoenzyme that oxidises
catalyses the production of the reactive N-acetyl- paracetamol. This conclusion arises largely from the
p-benzoquinone imine (NAPQI) that, as discussed observation that disulfiram, through its active me-
in section 4, is the cause of the characteristic cen- tabolite diethyldithiocarbamate, is a selective inhibi-
trilobular hepatotoxicity of overdoses of paraceta- tor of CYP2E1 and decreases the urinary excretion
mol. The reactive NAPQI reacts with glutathione to of the metabolites of NAPQI by about 70%.[4]
form the paracetamol-glutathione conjugate. The Paracetamol is also oxidised to NAPQI by sever-
bulk of this compound is converted to the cysteine al peroxidases, including myeloperoxidase[5,6] and

 2005 Adis Data Information BV. All rights reserved. Drug Safety 2005; 28 (3)
230 Graham et al.

the peroxidase function of COX-1.[7,8] A free radical The dimer and further polymers are produced selec-
species is also produced by these peroxidases lead- tively by peroxidases but, as far as we are aware,
ing to the final production of a paracetamol dimer their urinary output has not been examined. At this
and further polymers of paracetamol (figure 1). stage, the clinical significance of the oxidative me-
The main function of myeloperoxidase is to cat- tabolism of paracetamol by the peroxidases to reac-
alyse the formation of hypochlorous acid, an impor- tive metabolites is unknown, but it should be kept in
tant bactericidal product of neutrophils and mono- mind in any consideration of the toxicity of parac-
cytes. By acting as a substrate for myeloperoxidase, etamol.
paracetamol decreases the production of hypochlo-
rous acid. However, the decreased formation of hy- 3. Hepatotoxicity of Paracetamol
pochlorous acid after therapeutic doses of paraceta-
mol appears insufficient to produce a significant
3.1 Hepatotoxicity from Overdosage
attenuation of the bactericidal activity of these
cells.[9] The hepatotoxicity of overdoses of paracetamol
The conversion of paracetamol to reactive metab- is well known and is the most worrisome aspect of
olites by myeloperoxidase raises the question of its use. The typical pattern of toxicity is possible
their potential toxicity when produced by this en- nausea and vomiting in the first 12–24 hours. Be-
zyme. The metabolism of several drugs, such as tween 2 and 3 days after overdosage, plasma con-
propylthiouracil and clozapine,[10,11] by myeloper- centrations of the hepatic enzymes ALT and AST
oxidase in neutrophils and monocytes is associated begin to rise with a subsequent increase in bilirubin
with the development of agranulocytosis and sys- concentration and prolongation of prothrombin
temic lupus. There are also claims that paracetamol time. In severe cases, coma and bleeding develop.
very occasionally causes agranulocytosis in some Death often follows unless the patient receives a
patients,[12] although definitive evidence is still lack- liver transplant. The toxicity results from centrilobu-
ing. Furthermore, given its widespread use and the lar necrosis. Emergency treatment for this condition
very rare reports of possible associations between consists of gastric lavage and administration of N-
agranulocytosis and the intake of paracetamol, any acetylcysteine, which must be commenced within
causation of agranulocytosis by paracetamol must about the first 24 hours after overdosage, i.e. before
be very rare. centrilobular necrosis develops fully.[2]
The production of reactive metabolites through A significant clinical question in acute medical
the peroxidase function of COX-1 is also of note units concerns the identification of patients who
(figure 1).[7,8] It is presumed that COX-2 metabolises have, in fact, taken substantial doses of paracetamol.
paracetamol in a similar way to COX-1. Although Patients often supply incorrect or unreliable histo-
the COX isoenzymes are known only as cyclo- ries of their drug intake. Paracetamol is a widely
oxygenases, both have dual functions, namely used drug and it may be present when other drugs
cyclo-oxygenase and peroxidase activities. The me- have been taken in overdose. A nomogram relates
tabolism of paracetamol by the peroxidase activities the plasma concentrations to the time after dosage
of these enzymes may be highly relevant to its and indicates if the patient has taken a dangerous
mechanism of action.[3] Furthermore, excessive pro- overdose.[14] However, the nomogram is often diffi-
duction of reactive metabolites of paracetamol via cult to apply because the time of dosage is unknown
COX enzymes could result in cellular toxicity in the or multiple large doses have been taken. It follows
kidney medulla.[13] that N-acetylcysteine should be administered irre-
It should be emphasised that the extent of in vivo spective of the plasma concentrations if the bio-
metabolism of paracetamol by myeloperoxidase and chemical features (see section 3.3) are consistent
the peroxidase function of the COX isoenzymes are with an overdose and if an acute overdose or repeat-
unknown. It is widely considered that CYP2E1 is edly very high doses have possibly been taken.
the major source of NAPQI,[4] but the contribution Although questions remain about the details of
of the peroxidases in vivo has not been determined. the mechanism of paracetamol hepatotoxicity, it is

 2005 Adis Data Information BV. All rights reserved. Drug Safety 2005; 28 (3)
Tolerability of Paracetamol 231

widely accepted that the toxicity of paracetamol of paracetamol in humans.[4] Thus, it has been sug-
results from its oxidative metabolism to NAPQI. gested that the CYP2E1 inhibitor disulfiram may be
The commonly accepted mechanism is that NAPQI useful in decreasing the hepatotoxicity of paraceta-
reacts with glutathione and, when hepatocellular mol in some patients.[4] However, at present there is
glutathione is very much depleted by large amounts no definition of patients in whom disulfiram may be
of NAPQI, the NAPQI reacts with the thiol groups any more useful than the current effective treatment,
of liver proteins, which results in centrilobular ne- N-acetylcysteine. In vivo, N-acetylcysteine yields
crosis.[2] Depletion of glutathione may also provide cysteine, which is then conjugated with glycine and
an oxidant stress to cells or decrease the activities of glutamic acid to form reduced glutathione. Provided
enzymes for which glutathione is a cofactor.[2] Both that it is administered within about 24 hours after the
mechanisms have been cited in suggestions that overdosage, N-acetylcysteine provides a very effec-
lower than average levels of glutathione in the liver tive treatment for paracetamol overdose.[2]
or other tissues predispose to the toxicity induced by
paracetamol. Thus, it has been claimed that mal- 3.2 Hepatotoxicity from Therapeutic Doses
nourished patients or patients with hepatitis C, cir- in Adults?
rhosis or AIDS, who possibly have low levels of Following the discovery of the hepatotoxicity of
glutathione, should be at great risk of hepatotoxicity, overdoses of paracetamol, many cases of hepatotox-
even from therapeutic doses of paracetamol.[14,15] icity were reported with use of therapeutic doses of
However, recent reviewers have concluded that paracetamol. However, recent critical reviews in
there is no convincing evidence that patients with this field indicate that overdosage is the major cause
these disease states have any increased incidence or of the toxicity.[18-20]
severity of toxicity in the liver or other organs.[14,15] Careful reviews of hepatotoxicity in adults indi-
The pre-existing depletion of glutathione appears to cate that there are two types of liver changes associ-
be insufficient to potentiate hepatotoxicity from ated with alleged therapeutic doses of paraceta-
therapeutic doses of paracetamol.[15] However, it mol.[18-20] One type is consistent with overdose but
should be emphasised that depletion of glu- the second type may not be caused by paracetamol.
tathione leading to paracetamol-induced hepatotox- In the first type, affected patients have developed
icity should be considered a kinetic phenomenon centrilobular necrosis but the plasma concentrations
rather than a static situation. Hepatotoxicity devel- of paracetamol during hospitalisation indicate that
ops when the rate of production of the reactive overdoses were taken within the 2–3 days before
metabolite of paracetamol markedly exceeds the hospitalisation. Although therapeutic dosage has
rate of supply of reduced glutathione and toxicity been claimed, most patients have probably taken
occurs only when the glutathione concentrations in overdoses. An example of this first type of hepato-
the liver are considerably depleted.[15] toxicity was outlined by Bonkovsky et al.[19] A 67-
Further insight into the intracellular proteins that year-old man with a history of chronic cardiovascu-
mediate toxicity has been gained from examining lar, renal and pulmonary impairment developed the
the effects of targeted gene disruption on paraceta- classic features of paracetamol overdosage, includ-
mol toxicity in mice. Such studies have confirmed in ing elevated transaminase levels, a slow increase in
vivo that the CYP2E1 and CYP1A2 forms of CYP serum bilirubin level, and centrilobular hepatic ne-
are largely responsible for mediating toxicity in this crosis. There was also evidence of acute renal fail-
species.[16] Also, loss of the reduction/oxidation- ure. He claimed to have taken 1–3g of paracetamol
sensitive transcription factor Nrf2 enhances parac- daily for a short period of time, with the last dose
etamol toxicity by suppressing the induction of both having been taken 72 hours before hospitalisation.
paracetamol-detoxifying enzymes and glutathione The plasma concentration of paracetamol, presuma-
biosynthetic enzymes, resulting in enhanced parac- bly measured near the time of admission, was 27.5
etamol-mediated oxidative stress.[17] mg/L. After the patient had recovered from his
As outlined in section 2, CYP2E1 is the major hepatotoxicity, the elimination half-life of paraceta-
CYP responsible for oxidative hepatic metabolism mol was 7.9 hours. Based on this half-life, extrapo-

 2005 Adis Data Information BV. All rights reserved. Drug Safety 2005; 28 (3)
232 Graham et al.

lation of his plasma concentrations to 68 hours ment with paracetamol allowed the liver enzymes to
before the plasma sample (allowing for 4 hours to return to normal, but the enzyme concentrations in
attain peak plasma concentrations) yielded a plasma blood increased again after single doses of paraceta-
concentration of 11 g/L. This is an impossibly high mol.
value. The volume of distribution of paracetamol is The hepatic safety of paracetamol is also evident
approximately 50L and this body content would from recent studies. Firstly, dosage with 8 g/day
require a single dose of about 500g of paracetamol (twice the currently recommended maximal daily
or accumulation after multiple dosage to a body dosage) for 3 days did not increase hepatic tran-
content of this level. After recovery from the acute saminase levels in healthy subjects.[23] This high
hepatotoxicity, this patient had a low level of hepatic dosage is not recommended at this stage but indi-
glutathione and a prolonged elimination half-life of cates the safety of therapeutic dosages of paraceta-
paracetamol. These factors were suggested to indi- mol. Furthermore, the incidences of adverse hepatic
cate a greater than usual hepatic sensitivity to and renal events, which are not necessarily caused
paracetamol. However, the most likely reason for by paracetamol, are very similar at dosages of 3 g/
hepatotoxicity in this patient is still an acute over- day and 4 g/day of paracetamol in elderly pa-
dose of paracetamol over the preceding 2–3 days, as tients.[24]
Prescott[18] concluded from his survey of this and The overall conclusion is that hepatotoxicity
similar cases. In general, a patient’s history of parac- from therapeutic doses of paracetamol is most un-
etamol dosage cannot be accepted if the clinical and common.
biochemical changes observed, and particularly the
plasma concentrations of paracetamol measured, are 3.3 Hepatotoxicity from Therapeutic Doses
at great variance with the patient’s statements. A in Children?
number of recent reviews have emphasised this very
reasonable point.[14,15,18,20,21] Liver damage after paracetamol dosing for thera-
The second type of hepatoxicity from alleged peutic purposes has been widely reported in chil-
therapeutic dosage of paracetamol is, on examina- dren. The signs are very similar to those seen in
tion, found not to be centrilobular necrosis, the acute overdosage.[25] Furthermore, the plasma con-
classical hepatotoxic change seen with paracetamol. centrations of paracetamol, when they have been
A variety of syndromes, including chronic active measured, generally indicate that, as in adults, ad-
hepatitis, cholestasis and primary biliary cirrhosis, ministration of excessive doses is the probable cause
have developed during treatment with paracetamol of liver failure in most patients.[25,26] The drug may
but are generally not proven to be caused by have been administered for therapeutic purposes,
paracetamol treatment. As noted by Prescott,[18] “in but the parents or guardians have, in fact, adminis-
some cases, it is not clear why paracetamol should tered excessive doses. Therefore, possible overdos-
have been implicated as a cause of what appears to age with paracetamol should be considered in chil-
be naturally occurring liver disease”. dren who have a prodromal illness with en-
Idiosyncratic hepatic reactions after low doses of cephalopathy, even if the dosage has been stated to
paracetamol do occur but are extremely rare. In his be correct for the age or weight of the child.[25]
survey of hepatotoxicity in non-alcoholic adults, Investigations should include measurement of glu-
Prescott[18] found only one case in which challenge cose levels, prothrombin time, aminotransaminase
dosage with therapeutic doses of paracetamol yield- levels, bilirubin levels, acid-base status and parac-
ed convincing evidence of the signs of any hepatic etamol concentrations in plasma. As in adults, hy-
disease. In this case, the toxic change resembled poglycaemia, prolonged prothrombin time, high
chronic active hepatitis. More recently, two patients levels of aminotransaminases (>4 000 IU/L) and
with melanoma receiving interferon-α showed levels of bilirubin (<200 µmol/L) are very sugges-
marked increases in the plasma concentrations of tive of recent overdosage.[25]
ALT while taking relatively low dosages of parac- Although a nomogram is available to predict the
etamol (500–1500 mg/day).[22] Cessation of treat- likelihood of overdosage with single overdoses of

 2005 Adis Data Information BV. All rights reserved. Drug Safety 2005; 28 (3)
Tolerability of Paracetamol 233

paracetamol in adults, the potentially toxic levels of that “although the possibility remains that chronic
paracetamol are less clear after overdosage in chil- consumption of alcohol does increase the risk of
dren, especially after repeatedly excessive doses. paracetamol hepatotoxicity in man, there is insuffi-
Children should, therefore, be treated with N- cient evidence to support the alleged major toxic
acetylcysteine if biochemical indices are consistent interaction”.
with paracetamol-induced hepatotoxicity and there There are two major reasons for concluding that
is any evidence of the child receiving paracetamol. alcohol does not potentiate the hepatotoxicity of
Evidence of overdosage can be gathered from care- paracetamol sufficiently to make therapeutic doses
ful discussions with the parents or guardians and by hepatotoxic. Firstly, as with cases of hepatotoxicity
measuring the plasma concentrations of the drug or ascribed to therapeutic doses of paracetamol in non-
detecting it in the urine. Paracetamol is, of course, a alcohol drinking adults and children, extrapolation
widely used drug and liver disease in many children of the plasma concentrations in moderate and
will be unrelated to use of the drug. If the liver heavy alcohol drinkers with hepatotoxicity indicates
function tests remain abnormal, exclusion of other that excessive doses have often been adminis-
causes of liver failure is important in the long-term tered.[14,20,21] Secondly, all prospective studies indi-
management of these patients. Other possible causes cate that therapeutic doses of paracetamol are not
include hepatotoxicity from other drugs, hepatitis A, hepatotoxic in alcoholic patients.[14,20,21] However, a
B or C viruses, Epstein-Barr virus, cytomegalovirus limitation in the prospective studies should be noted.
or inborn errors of metabolism, such as Wilson’s This is that the dosage of paracetamol has generally
disease and α1-antitrypsin deficiency.[25] However, been short-term (from 1 to 5 days). Although diffi-
tests for these diseases are generally too slow for the cult, it would be useful to conduct longer term
emergency management of patients who have possi- prospective studies.
bly taken overdoses of paracetamol and the antidote, The possible effect of alcohol on the hepatotoxic-
N-acetylcysteine, should therefore be started before ity of paracetamol is being investigated at the level
these diseases can be eliminated. of the metabolic interaction between the two com-
3.4 Hepatotoxicity from Therapeutic Doses in
pounds. Alcohol has variable, although generally
Patients Drinking Alcohol?
modest, effects on the CYP2E1 pathway, which is
largely responsible for production of the hepatotoxic
It has been claimed widely that alcohol poten- metabolites of paracetamol.[4] Alcohol induces this
tiates the hepatotoxicity of paracetamol sufficiently enzyme, but it also inhibits CYP2E1 while the alco-
to make therapeutic doses potentially hepatotox- hol remains in the body. Alcohol may, therefore,
ic.[20,21,27] This claim has been formalised with the protect the liver by inhibiting the oxidative metabo-
term ‘alcohol-paracetamol syndrome’ to describe lism of paracetamol. Subsequently, alcohol may
the hepatotoxicity that is said to occur from the make the liver more sensitive to paracetamol be-
ingestion of therapeutic doses of paracetamol in cause hepatic concentrations of CYP2E1 appear to
moderate to heavy drinkers of alcohol.[27] This belief decline at a slower rate than the plasma concentra-
is also reflected in a ruling by the US FDA, which tions of alcohol. Thus, alcohol may be removed
now requires that all packages of paracetamol sold from the body and unable to inhibit the metabolism
in the US should be labelled with the warning “if of paracetamol while the concentrations of CYP2E1
you consume 3 or more alcoholic drinks every day, are still elevated and capable of a higher than normal
you should ask your doctor whether you should take rate of conversion of paracetamol to hepatotoxic
acetaminophen (paracetamol) or other pain reliev- metabolites. However, alcohol appears to produce
ers/fever reducers. Acetaminophen may cause liver only a small increase (range from 2% to 38%) in the
failure” (see section 4). However, once again, criti- oxidative metabolism of paracetamol when the
cal examination indicates that many cases of paracetamol is administered 8 hours after the cessa-
hepatoxicity in alcohol drinkers have been due to the tion of an infusion of alcohol sufficient to produce a
ingestion of overdoses, not therapeutic doses, of blood concentration of 0.1g per 100mL.[28] Never-
paracetamol.[14,20,21] Indeed, Prescott[21] concluded theless, simulations indicate that up to an approxi-

 2005 Adis Data Information BV. All rights reserved. Drug Safety 2005; 28 (3)
234 Graham et al.

mate doubling of the rate of formation of the reac- 3.5 Use of Paracetamol in Alcoholics and in
tive metabolite, NAPQI, could occur after cessation Patients with Liver Diseases
of an intake of alcohol sufficient to produce a blood
Limited clinical studies indicate that paracetamol
concentration of 0.3g per 100mL.[28] More directly,
can be administered to patients with chronic liver
testing indicates that 1g of paracetamol four times diseases.[35-37] Paracetamol does not exacerbate sta-
per day does not lead to biochemical evidence of ble chronic liver disease,[35] for example. Further-
liver damage in alcoholic patients in the period just more, the metabolism of paracetamol appears nor-
after cessation of alcohol use, the point at which mal in patients with liver disease although the elimi-
they should be maximally sensitive to paraceta- nation half-life is prolonged by an average of 75% in
mol.[29] patients with severe liver disease.[36] In an authorita-
Calculations of the rate of formation of NAPQI tive textbook, paracetamol (up to 1g three times per
day) is recommended as the optimal analgesic in
and glutathione do not indicate that paracetamol
patients with chronic liver disease.[37] However, it is
should be markedly more toxic in alcoholics than in prudent to monitor liver function in such patients
non-alcoholics.[14,15] Furthermore, paracetamol (4 g/ and use of paracetamol in patients with hepatic
day for 3 days) does not reduce glutathione levels in diseases should be kept as short as possible.
the plasma of alcoholics.[30] Glutathione in the liver What alternative analgesics or antipyretics can be
may be depleted in chronic alcoholics, but this does used in patients with liver disease or in alcoholics?
not appear to increase the risk of paracetamol-in- The NSAIDs, such as aspirin or ibuprofen, are rela-
duced hepatotoxicity, possibly because the oxida- tively contraindicated because of the morbidity and
tive metabolism of paracetamol is also inhibited in mortality resulting from gastrointestinal damage
these patients.[14,15] However, the interaction be- produced by NSAIDs and the risk of bleeding in
tween alcohol and paracetamol is still a contentious patients with varices.[38] Alcohol tends to potentiate
area and it is reasonable to suggest that alcoholics the gastric damage produced by non-selective
NSAIDs[39] and initial bleeding from varices is also
who have taken an overdose of paracetamol
increased.[38] The selective COX-2 inhibitors, such
should be treated with N-acetylcysteine at lower as celecoxib, may ultimately prove to be useful, but
plasma concentrations of paracetamol than non-al- clinical evidence for the safety of these drugs com-
coholics.[31] bined with alcohol and in liver failure is lacking at
Overdoses of paracetamol may be more common present. Opioid analgesics may be used in severe
in alcoholics than in the remainder of the popula- pain, but care should be taken with the dosage of
tion,[32] but this does not necessarily mean that these agents because of possible decreased metabol-
chronic use of alcohol potentiates the hepatotoxicity ic clearance and respiratory depression.
of therapeutic doses of paracetamol. Rather, the Overall, it appears that paracetamol is a reasona-
ble analgesic or antipyretic drug to use in alcoholics
depression and other psychological and psychiatric
and patients with liver diseases. However, the dos-
factors associated with alcoholism may make them age should be monitored and the number of tablets
more likely to take an overdose of paracetamol.[33,34] or capsules of paracetamol available for alcoholic
Furthermore, the memory loss often associated with patients should be restricted in order to reduce the
severe alcoholism may make affected individuals chances of overdosage. In this regard, it is notable
unaware of having taken excessive doses. that very large packs of paracetamol tablets are
Overall, it appears unlikely that alcohol increases marketed in the US despite the warning on the label.
the hepatotoxicity of paracetamol sufficiently to
cause toxicity during therapeutic dosages. However, 4. Labelling and Packaging
the toxicity of paracetamol overdoses may be in- No warnings about the use of paracetamol with
creased in alcoholics. This is an area of ongoing alcohol are specified in most countries but, as out-
research. lined in section 3.4, paracetamol sold in the US is

 2005 Adis Data Information BV. All rights reserved. Drug Safety 2005; 28 (3)
Tolerability of Paracetamol 235

accompanied by a warning that it must be taken with the US. It is very difficult to evaluate the effect of
care if alcohol is also taken in moderate to excessive restricting the size of packages of any drug. Howev-
amounts. er, preliminary evidence suggests that the numbers
What are the reasons for the FDA labelling of of paracetamol overdoses have declined since the
paracetamol? There are two possible reasons for pack size was reduced in the UK,[45] despite the fact
inclusion of this warning about the potential for that it is still easy to obtain several packs of
hepatotoxicity of paracetamol in alcoholics. The paracetamol tablets or capsules.
first is that critical analyses of reports of hepatotox-
icity said to be produced by therapeutic doses of 5. Gastrointestinal Tolerability
paracetamol in alcoholics have only recently been In pharmacology texts and reviews, it is frequent-
published. The second reason is that the FDA may ly stated that paracetamol has excellent gastrointes-
have tried to err on the side of patient safety. This tinal tolerability. These statements are based on
was a reason admitted to by the American College of prospective studies that show that paracetamol does
Rheumatology (ACR) in relation to its 2000 Guide- not damage the gastrointestinal tract.[46-48] As dis-
lines for the Medical Management of Osteoarthri- cussed in the introduction, this tolerability has been
tis.[40] In these guidelines, the ACR recommended related to selective inhibition of cellular pros-
that paracetamol should be avoided in patients with taglandin synthesis involving COX-2 by paraceta-
chronic alcohol abuse and used with caution in mol, although further work is required to support
patients with existing liver disease. Subsequent cor- this hypothesis concerning its mechanism of ac-
respondence in Arthritis and Rheumatism[41,42] was tion.[3] The favourable gastrointestinal tolerability of
generally critical of these statements and, in reply, paracetamol is in contrast with that of aspirin, which
the spokesman for the ACR[43] stated that it was is associated with considerable risk of bleeding even
“… better to err on the side of patient safety given when given at low doses.[49]
that alternative treatments are available …”. While The excellent gastrointestinal tolerability of
statements on product labels or in information given paracetamol has been confirmed in recent meta-
to physicians about the possible hepatotoxicity of analyses of case-control studies. Lewis et al.[50]
therapeutic dosages of paracetamol are currently found no significant effect of paracetamol on the
being considered in countries other than the US, it is gastrointestinal tract at any dose. Another researcher
important to note that poorly justified statements are (Henry DA, personal communication) found that
not helpful and decrease attention to well accepted paracetamol had a significantly increased relative
warnings about other drugs. risk of upper gastrointestinal reactions (perforations,
It is also important that the correct dosage of ulcers or bleeding), but the pooled relative risk was
paracetamol should be emphasised to patients. How- only 1.5 (95% CI 1.21, 1.85). These figures were
ever, the potential for death from overdoses of calculated from the results of ten studies in which
paracetamol should still not be widely broadcast some adjustment was made for potential con-
because of the danger of popularising overdosage founding variables such as alcohol and other
with this drug. There is now documented evidence NSAIDs. Although statistically significant, the asso-
that imitations of suicides occur following news ciation is weak and very much within the area where
items, films and television dramas in which suicides confounding variables make the risk uncertain.
are reported or depicted. Not all news reports or Three epidemiological studies have indicated
fictional stories have been clearly followed by sui- a greater incidence of adverse gastrointestinal
cides or attempted suicides, but many have been.[44] reactions with increasing daily doses of paraceta-
Judging from the occurrence of copycat suicides, the mol.[51-53] In one of these studies, patients taking
label on US packs about the hepatotoxicity of paracetamol had a higher incidence of gastrointesti-
paracetamol may encourage some patients to take nal events than patients taking NSAIDs, but this was
overdoses of the drug. related to older age and risk factors for gastropathy
In recent years, the package size of paracetamol in the patients taking paracetamol.[53] In the same
has been limited in many countries, although not in study, dyspepsia was far more common than ulcers

 2005 Adis Data Information BV. All rights reserved. Drug Safety 2005; 28 (3)
236 Graham et al.

or upper gastrointestinal bleeding in the groups tak- 6. Renal Tolerability

ing the higher doses of paracetamol.[53] This was
described as “somewhat reassuring”, although the Non-selective NSAIDs and selective COX-2 in-
dyspepsia is still troublesome and often causes cost- hibitors may precipitate acute renal failure in pa-
ly medical investigation. tients with risk factors such as congestive cardiac
The relationship between increasing dose of failure, pre-existing renal impairment and trans-
paracetamol and gastrointestinal events reported in planted kidneys.[55] By contrast, exacerbation of car-
these studies[51-53] may be a biased result. It has been diac failure has not been reported with paracetamol
suggested that the higher incidence of bleeding in and acute renal failure has not been associated with
patients taking the larger doses of paracetamol could therapeutic doses of the drug.[55] However, over-
occur because “physicians would be especially care- doses of paracetamol may produce acute renal fail-
ful not to prescribe high doses of NSAIDs to those at ure as a result of acute tubular necrosis, possibly
greatest risk of GI bleeding” and would therefore because of metabolism of the drug to reactive me-
use paracetamol instead.[1] The gastrointestinal safe- tabolites by the peroxidase function of COX-1 or
ty of paracetamol is widely advertised and patients COX-2.[13]
who know or suspect they have upper gastrointesti- Possibly through their renal effects, the non-
nal disease may have bought over-the-counter prep- selective NSAIDs and the selective COX-2 inhibi-
arations of paracetamol. Therefore, a greater inci- tors may decrease the efficacy of diuretics and an-
dence of gastrointestinal events overall, and particu- tihypertensives.[55] No such interactions have been
larly with high doses of paracetamol, may be related to use of paracetamol although NSAIDs and
expected. paracetamol were associated with the development
Protopathic bias may also occur. For paraceta- of hypertension in young women (aged from 25 to
mol, this can occur because this agent may be used 42 years) in one cohort study.[56]
to treat the pain or discomfort of early gastrointesti- Studies of the effects of paracetamol on the ex-
nal disease.[1] Peptic ulcer is then associated with the cretion of prostaglandins and sodium have yielded
use of paracetamol. An indication of this bias with inconsistent results.[55] Although the excretion of
paracetamol is that an association between the use of prostaglandins, their metabolites and sodium have
paracetamol and gastrointestinal bleeding was seen been reduced in most studies,[55] this has not been a
when paracetamol was used for indigestion but not universal result.[55] Overall, the effects of paraceta-
when it was used for headaches or colds.[54] The mol on the renal output of prostaglandins and sodi-
results of case control or cohort studies are corrected um appear to be weaker than those of the non-
for risk factors, but it is difficult to remove con- selective and selective COX-2 inhibitors.[55] Howev-
founding variables completely when the reasons for er, direct comparative studies are required, particu-
the use of drugs are not known. This is a particular larly in patients with risk factors such as cardiac
problem with widely available drugs, such as failure, for the retention of sodium and water. Inves-
paracetamol. tigational studies of sodium clearance should also be
Because of the lack of acute changes in the upper conducted on the first day of treatment with
gastrointestinal tract observed with use of paraceta- paracetamol because it is at this time that NSAIDs
mol, the majority view is still that this drug is free of show their greatest inhibitory effect on this aspect of
major gastrointestinal toxicity. Nevertheless, the renal function.[55]
questions raised by epidemiological studies remain The renal toxicity of paracetamol has mainly
and may not be answered finally until a large scale been examined in terms of development of chronic
randomised prospective clinical trial of major renal failure, particularly that resulting from analge-
events, such as perforations, ulcer or bleeding, is sic nephropathy.[57] The results have been inconsis-
conducted in paracetamol users. Data from such a tent. A recent reviewer concluded that “there is
trial might also indicate if paracetamol causes dys- currently insufficient evidence to conclude that the
pepsia. At this stage, there is no clear evidence habitual use of paracetamol is associated with an
that it does. increased risk of chronic renal disease”.[57] Despite

 2005 Adis Data Information BV. All rights reserved. Drug Safety 2005; 28 (3)
Tolerability of Paracetamol 237

this conclusion, there are studies showing such an mended in patients taking paracetamol and warfarin,
association, most recently that by Fored et al.[58] As particularly when regular and daily dosage with >2g
in other settings, it is difficult to remove con- paracetamol is started or stopped.[62,63]
founding variables in epidemiological studies of the
renal effects of paracetamol. In recent years at least, 8. Hypersensitivity Reactions
paracetamol has been considered to be safer for The metabolism of paracetamol to reactive com-
kidneys than the non-selective NSAIDs. Therefore, pounds suggests that it may act as a hapten and lead
as is the case with the use of paracetamol in other to hypersensitivity reactions. Although the mecha-
disease states, the prescription of paracetamol for nisms involved are unclear, such reactions do occur
patients with renal disease may have led to the with paracetamol but are very rare. For example,
association between paracetamol and renal disease. while allergic skin reactions to paracetamol are ex-
However, there may not be any causation underly- tremely rare, urticaria has been produced after ad-
ing the association. ministration of test oral doses.[64] Many patients who
have skin reactions to aspirin or other NSAIDs do
7. Haemostasis
not react to paracetamol.[64] Conversely, paraceta-
Therapeutic doses of paracetamol decrease the mol produces urticaria in a few patients who tolerate
synthesis of thromboxane A2 by platelets (a aspirin.[65] Anaphylactic shock also occurs very
COX-1-dependent system) after therapeutic dos- rarely in patients taking paracetamol,[66] although, in
age.[59] However, an almost total blockade of some cases, it is caused by an additive in the tablet
thromboxane A2 synthesis is required before platelet and not by the paracetamol itself.[67]
aggregation is markedly inhibited and therapeutic Despite the very low incidence of paracetamol-
doses of paracetamol therefore have no significant precipitated asthma, use of paracetamol was associ-
effect on platelet aggregation. This result contrasts ated with asthma in a controversial case control
with the marked anti-platelet effects of the non- study.[68] This found that the odds ratio for the
selective NSAIDs at therapeutic doses.[2] incidence of asthma increased with more frequent
Thrombocytopenia is also associated with use of the drug. Avoidance of aspirin was consid-
hepatotoxicity of paracetamol in a small proportion ered in this study, but the use of other non-selective
of overdoses[60] but is extremely rare at therapeutic NSAIDs was not. Because of this and other potential
doses. Nevertheless, immune thrombocytopenia has confounding variables, there is no definite proof that
been reported in a very small number of patients.[61] paracetamol increases the incidence of asthma.
This immune-induced thrombocytopenia may be in- Of more significance is the effect of paracetamol
duced by paracetamol sulfate, a major metabolite of in aspirin-induced asthma. Acute asthma is precipi-
paracetamol. Other hypersensitivity reactions of tated by aspirin and the non-selective NSAIDs in
paracetamol are extremely rare and are discussed in some asthmatics. The recorded proportion of
section 8. asthmatics who are sensitive to aspirin and the non-
Because of its weak effect on platelets, paraceta- selective NSAIDs is very variable but probably in
mol is considered safe in patients with clotting dis- the range of 4% to 20%.[69-71] These patients fre-
orders and in patients taking anticoagulants. Howev- quently develop nasal polyps before their sensitivity
er, paracetamol may increase the effect of warfarin to aspirin and the non-selective NSAIDs manifests
in some patients.[62,63] This potentiation of the effect itself. Paracetamol is well tolerated by most, but not
of warfarin is a contentious issue, but an interaction all, of these asthmatics.[69-71] Furthermore, the asth-
between paracetamol and warfarin is now included matic reaction is milder after exposure to paraceta-
in standard references on drug interactions.[62,63] mol than after dosage with the non-selective
Paracetamol is still recommended as a reasonable NSAIDs.[69-71] Unfortunately, detecting the few pa-
analgesic and antipyretic to use with warfarin be- tients with cross-reactivity between paracetamol and
cause of the weak effect of paracetamol on platelet the non-selective NSAIDs can only be achieved
aggregation at therapeutic dosages.[62,63] Neverthe- with provocation tests.[71] The general safety of
less, monitoring of the prothrombin time is recom- paracetamol in asthmatics has led to the recommen-

 2005 Adis Data Information BV. All rights reserved. Drug Safety 2005; 28 (3)
238 Graham et al.

dation that “routine warnings about paracetamol use (see section 7). The non-selective NSAIDs inhibit
in asthma are, therefore, not warranted”,[71] although COX-1 in platelets and may increase blood loss
medical personnel should be aware of this occasion- associated with childbirth or produce bleeding in the
al problem in patients in whom testing for the reac- baby.[75] By contrast, paracetamol should not.
tion or withdrawal of paracetamol is warranted.[70,71]
The low incidence of asthmatic reactions to parac- 11. Conclusions
etamol may be related to some inhibition of pros-
taglandin synthesis by COX-1. The conclusion is The major problem arising from the widespread
reached that non-selective NSAIDs all precipitate use of paracetamol is the ability of overdoses to
asthma in aspirin-sensitive asthmatics, whereas the cause hepatotoxicity as a result of metabolism of the
COX-2 selective asthmatics do not appear to pro- drug to reactive compounds. Many cases of
duce this syndrome.[72] hepatotoxicity have been claimed to be associated
with therapeutic doses of paracetamol, but critical
9. Occurrence of Cancers analysis indicates that many of these cases have
resulted from overdoses. The dosage of paraceta-
While elevated risks for the development of some mol, particularly in children and alcoholics, should
tumours of the urinary tract with use of paracetamol be controlled carefully in order to prevent
have been detected in some case control studies,[73] a hepatotoxicity. Medical practitioners, nurses and
recent reviewer concluded that “results do not sup- pharmacists should emphasise to patients the need to
port a major role for paracetamol in the development take the correct dosage of paracetamol, although
of cancers”.[73] Conversely, there are inconsistent discussion of the dangers of overdose in the media
reports of paracetamol decreasing the risk of ovarian should be restricted because of the risk of increasing
cancer[73,74] although, again, the results are insuffi- the numbers of suicides from overdoses. Adequate
cient to allow any definitive conclusions. labelling based on critical analysis of the literature is
also important.
10. Tolerability During Pregnancy Epidemiological reports of associations between
paracetamol and chronic renal disease, gastrointesti-
Paracetamol is preferred to the non-selective
nal damage and asthma may be biased, at least in
NSAIDs as an analgesic during pregnancy.[75] Stud-
part, because of the perceived safety of paracetamol
ies in mice show that prostaglandins formed through
in patients with these known diseases or because of
the COX-2 pathway are important in the early
other confounding variables. However, continuing
processes of pregnancy, including ovulation, fer-
surveillance of such possible associations is war-
tilisation and implantation.[76] COX-2 activity is also
ranted.
significant in the initiation of labour.[77] By inhib-
iting the COX-2 pathways in intact cells, the non-
Acknowledgements
selective NSAIDs, COX-2 selective inhibitors and
paracetamol could affect all of these processes. The The authors gratefully acknowledge discussions with Pro-
non-selective NSAIDs and the selective COX-2 in- fessor David Henry, Dr Richard Robson, Dr Robert Graham
hibitors delay labour[77,78] but, as far as we are and Dr Bridin Murnion in the preparation of this article. A
research project of Professor Graham has been supported by
aware, no effect of paracetamol on late pregnancy in
GlaxoSmithKline and Professor Day is a member of advisory
women has been reported. A recent cohort study boards for the companies marketing celecoxib (Pfizer and
indicated that NSAIDs, but not paracetamol, may Pharmacia), rofecoxib (Merck) and paracetamol (acetami-
increase the chance of miscarriage.[79] Prospective nophen) [GlaxoSmithKline]. Dr Scott is supported by the
studies of the NSAIDs and paracetamol on the early National Health and Medical Research Council of Australia
(Grant no. 222870).
processes of pregnancy and its maintenance are
clearly required.
A clear difference between the actions of References
1. Signorello LB, McLaughlin JK, Lipworth L, et al. Confounding
paracetamol and the non-selective NSAIDs is the by indication in epidemiological studies of commonly used
much weaker antiplatelet effect of paracetamol analgesics. Am J Ther 2002; 9: 199-204

 2005 Adis Data Information BV. All rights reserved. Drug Safety 2005; 28 (3)
Tolerability of Paracetamol 239

2. Roberts LJ, Morrow JD. Analgesic-antipyretic and antiinflam- 23. Gelotte CK, Auiler JF, Temple AR, et al. Clinical features of a
matory agents and drugs employed in the treatment of gout. In: repeat-dose multiple-day pharmacokinetics trial of acetami-
Goodman and Gilman’s The Pharmacological Basis of Thera- nophen at 4, 6, and 8 g/day [abstract]. J Toxicol Clin Toxicol
peutics. 10th ed. New York: McGraw-Hill, 2001: 687-731 2003; 41: 726
3. Graham GG, Robins S-A, Bryant KJ, et al. Inhibition of pros- 24. Ganry H, Pruvot F, Vesque D, et al. Liver and renal tolerance to
taglandin synthesis in intact cells by paracetamol (acetami- paracetamol: 3 g or 4 g per day? [in French]. Presse Med 2001;
nophen). Inflammopharmacology 2001; 9: 131-42 30: 724-38
4. Manyike PT, Kharasch ED, Kalhorn TF, et al. Contribution of 25. Miles FK, Kamath R, Dorney SF, et al. Accidental paracetamol
CYP2E1 and CYP3A to acetaminophen reactive metabolite overdosing and fulminant hepatic failure in children. Med J
formation. Clin Pharmacol Ther 2000; 67: 275-82 Aust 1999; 171: 472-5
5. O’Brien PJ, Khan S, Jatoe SD. Formation of biological reactive 26. Penna A, Buchanan N. Paracetamol poisoning in children and
intermediates by peroxidases: halide mediated acetaminophen hepatotoxicity. Br J Clin Pharmacol 1991; 32: 143-9
oxidation and cytotoxicity. Adv Exp Med Biol 1990; 283: 27. Draganov P, Durrence H, Cox C, et al. Alcohol-acetaminophen
51-64 syndrome. Postgrad Med 2000; 107: 189-95
6. Graham GG, Milligan MK, Day RO, et al. Therapeutic consid-
erations from pharmacokinetics and metabolism: ibuprofen 28. Thummel KE, Slattery JT, Ro H, et al. Ethanol and production
and paracetamol. In: Rainsford KD, Powanda MC, editors. of the hepatotoxic metabolite of acetaminophen in healthy
Safety and efficacy of non-prescription (over-the-counter) adults. Clin Pharmacol Ther 2000; 67: 591-9
analgesics and NSAIDs. Edinburgh: Kluwer Press, 1997: 29. Kuffner EK, Dart RC, Bogdan GM, et al. Effect of maximal
77-92 daily doses of acetaminophen on the liver of alcoholic patients:
7. Potter DW, Hinson JA. The 1- and 2-electron oxidation of a randomized, double-blind, placebo-controlled trial. Arch In-
acetaminophen catalyzed by prostaglandin H synthase. J Biol tern Med 2001; 161: 2247-52
Chem 1987; 262: 974-80 30. Palmer RB, Green JL, Kuffner EK, et al. Plasma reduced
8. Harvison PJ, Egan RW, Gale RH, et al. Acetaminophen and glutathione (GSH) in alcoholics receiving maximal therapeutic
analogs as cosubstrates and inhibitors of prostaglandin H doses of acetaminophen [abstract]. J Toxicol Clin Toxicol
synthase. Chem Biol Interact 1988; 64: 251-66 2003; 41: 726
9. Graham GG, Day RO, Milligan MK, et al. Current concepts of 31. Buckley NA, Srinivasan J. Should a lower treatment line be
the actions of paracetamol (acetaminophen) and NSAIDs. In- used when treating paracetamol poisoning in patients with
flammopharmacology 1999; 7: 255-64 chronic alcoholism?: a case for. Drug Saf 2002; 25: 619-24
10. Uetrecht JP. Idiosyncratic drug reactions: possible role of reac- 32. McMahon GT, McGarry K. Deliberate self-poisoning in an Irish
tive metabolites generated by leukocytes. Pharm Res 1989; 6: county hospital. Irish J Med Sci 2001; 170: 94-7
265-73 33. Hufford MR. Alcohol and suicidal behaviour. Clin Psychol Rev
11. Uetrecht JP. Metabolism of clozapine by neutrophils: possible 2001; 21: 797-811
implications for clozapine-induced agranulocytosis. Drug Saf 34. Thase ME, Salloum IM, Cornelius JD. Comorbid alcoholism
1992; 7 Suppl. 1: 51-6 and depression: treatment issues. J Clin Psychiatry 2001; 62
12. Lacotte J, Perrin C, Mosquet B, et al. Agranulocytosis from Suppl. 20: 32-41
paracetamol. Therapie 1990; 45: 438-9 35. Benson GD. Acetaminophen in chronic liver disease. Clin
13. Mohandas J, Duggin GG, Horvath JS, et al. Metabolic oxidation Pharmacol Ther 1983; 33: 95-101
of acetaminophen (paracetamol) mediated by cytochrome P- 36. Forrest JA, Adriaenssens P, Finlayson ND, et al. Paracetamol
450 mixed-function oxidase and prostaglandin endoperoxide metabolism in chronic liver disease. Eur J Clin Pharmacol
synthetase in rabbit kidney. Toxicol Appl Pharmacol 1981; 61: 1979; 15: 427-31
252-9
37. McIntyre N. The general management of liver disease. In:
14. Rumack BH. Acetaminophen hepatotoxicity: the first 35 years. J
Bircher J, Benhamou J-P, McIntyre N, et al., editors. Oxford
Toxicol Clin Toxicol 2002; 40: 3-20
textbook of clinical hepatology. Oxford: Oxford University
15. Lauterburg BH. Analgesics and glutathione. Am J Ther 2002; 9: Press, 1999: 1917-23
225-33
16. Gonzalez FJ. The use of gene knockout mice to unravel the 38. de Ledinghen V, Heresbach D, Fourdan O, et al. Anti-inflam-
mechanisms of toxicity and chemical carcinogenesis. Toxicol matory drugs and variceal bleeding: a case control study. Gut
Lett 2001; 120: 199-208 1999; 44: 270-3
17. Chan K, Han X-D, Kan YW. An important function of Nrf2 in 39. Lanza FL, Royer GL, Nelson RS, et al. Ethanol, aspirin, and the
combating oxidative stress: detoxification of acetaminophen. gastroduodenal mucosa: an endoscopic assessment. Am J Gas-
Proc Natl Acad Sci U S A 2001; 98: 4611-6 troenterol 1985; 80: 767-9
18. Prescott LF. Therapeutic misadventure with paracetamol: fact or 40. American College of Rheumatology Subcommittee on Osteoar-
fiction. Am J Ther 2000; 7: 99-114 thritis Guidelines. Recommendations for the medical manage-
19. Bonkovsky HL, Kane RE, Jones DP, et al. Acute hepatic and ment of hip and knee osteoarthritis. 2000 update. Arthritis
renal toxicity from low doses of acetaminophen in the absence Rheum 2000; 43: 1905-15
of alcohol abuse or malnutrition: evidence for increased sus- 41. Lynch JM, Fox TM. Use of acetaminophen in the treatment of
ceptibility to drug toxicity due to cardiopulmonary and renal osteoarthritis in patients with liver disease: comment on the
insufficiency. Hepatology 1994; 19: 1141-8 2000 update of the American College of Rheumatology recom-
20. Dart RC, Kuffner EK, Rumack BH. Treatment of pain or fever mendations for the management of hip and knee osteoarthritis
with paracetamol (acetaminophen) in the alcoholic patient: a [letter]. Arthritis Rheum 2001; 44: 2448-9
systematic review. Am J Ther 2000; 7: 123-34 42. Dart RC, Kuffner EK. Use of acetaminophen in alcoholic pa-
21. Prescott LF. Paracetamol, alcohol and the liver. Br J Clin tients: comment on the 2000 update of the American College
Pharmacol 2000; 49: 291-301 of Rheumatology recommendations for the management of
22. Fabris P, Palma MD, de Lalla F. Idiosyncratic acute hepatitis hip and knee osteoarthritis [letter]. Arthritis Rheum 2001;
caused by paracetamol in two patients with melanoma treated 44: 2449
with high-dose interferon-α [letter]. Ann Intern Med 2001; 43. Hochberg MC. Reply to letters. Arthritis Rheum 2001; 44:
134: 345 2455-6

 2005 Adis Data Information BV. All rights reserved. Drug Safety 2005; 28 (3)
 240 Graham et al.

44. Gould MS. Suicide and the media. Ann NY Acad Sci 2001; 932: 62. Stockley IH. Drug interactions. 3rd ed. Oxford: Blackwell, 1994
200-24
63. Hansten PD, Horn JR. Drug interactions. St Louis (MO): Facts
45. Langford NJ, Aruna RS, Mutimer D, et al. The impact of pack and Comparisons Publishing Group, 2003
size legislation on paracetamol (acetaminophen) poisoning in
the West Midlands (United Kingdom). J Toxicol Clin Toxicol 64. Quaratino D, Romano A, Papa G, et al. Long-term tolerability of
2003; 41: 419-20 nimesulide and acetaminophen in non-steroidal anti-inflam-
matory drug-intolerant patients. Ann Allergy Asthma Immu-
46. Ivey KJ, Settree P. Effect of paracetamol (acetaminophen) on nol 1997; 79: 47-50
gastric ionic fluxes and potential difference in man. Gut 1976;
17: 916-9 65. De Paramo BJ, Gancedo SQ, Cuevas M, et al. Paracetamol
(acetaminophen) hypersensitivity. Ann Allergy Asthma Im-
47. Hoftiezer JW, O’Laughlin JC, Ivey KJ. Effects of 24 hours of munol 2000; 85: 508-11
aspirin, Bufferin, paracetamol and placebo on normal gas-
troduodenal mucosa. Gut 1982; 23: 692-7 66. Kvedariene V, Bencherioua AM, Messaad D, et al. The accura-
cy of the diagnosis of suspected paracetamol (acetaminophen)
48. Lanza FL, Codispoti JR, Nelson EB. An endoscopic comparison hypersensitivity: results of a single-blinded trial. Clin Exp
of gastroduodenal injury with over-the-counter doses of Allergy 2002; 32: 1366-9
ketoprofen and acetaminophen. Am J Gastroenterol 1998; 93:
1051-4 67. Ronnau AC, Wulferink M, Gleichmann E, et al. Anaphylaxis to
polyvinylpyrrolidine in an analgesic preparation. Br J Derma-
49. Weil J, Colin-Jones D, Langman M, et al. Prophylactic aspirin tol 2000; 143: 1055-8
and risk of peptic ulcer bleeding. BMJ 1995; 310: 827-30
68. Shaheen SO, Sterne JA, Songhurst CE, et al. Frequent
50. Lewis SC, Langman MJS, Laporte JR, et al. Dose-response paracetamol use and asthma in adults. Thorax 2000; 55: 266-7
relationships between individual nonaspirin nonsteroidal anti-
inflammatory drugs (NANSAIDs) and on serious upper gas- 69. Morassut P, Yang W, Karsh J. Aspirin intolerance. Semin
trointestinal bleeding: a meta analysis based on individual Arthritis Rheum 1989; 19: 22-30
patient data. Br J Clin Pharmacol 2002; 54: 320-6
70. Jenkins C. Recommending analgesics for people with asthma.
51. Savage RL, Moller PW, Ballantyne CL, et al. Variation in the Am J Ther 2000; 7: 55-61
risk of peptic ulcer complications with nonsteroidal antiin-
71. Levy S, Volans G. The use of analgesics in patients with asthma.
flammatory drug therapy. Arthritis Rheum 1993; 36: 84-90
Drug Saf 2001; 24: 829-41
52. Garcia Rodriguez LA, Hernandez-Diaz S. The risk of upper
gastrointestinal complications associated with nonsteroidal an- 72. West PM, Fernandez C. Safety of COX-2 inhibitors in asthma
ti-inflammatory drugs, glucocorticoids, acetaminophen, and patients with aspirin hypersensitivity. Ann Pharmacother
combinations of these agents. Arthritis Res 2001; 3: 98-101 2003; 37: 1497-501

53. Rahme E, Pettitt D, LeLorier J. Determinants and sequelae 73. Friis S, Nielsen GL, Mellemkjaer L, et al. Cancer risk in persons
associated with utilization of acetaminophen versus traditional receiving prescriptions for paracetamol: a Danish cohort study.
nonsteroidal antiinflammatory drugs in an elderly population. Int J Cancer 2002; 97: 96-101
Arthritis Rheum 2002; 46: 3046-54 74. Cramer DW, Harlow BL, Titus-Ernstoff L, et al. Over-the-
54. Langman MJ, Coggon D, Spiegelhalter D. Analgesic intake and counter analgesics and the risk of ovarian cancer. Lancet 1998;
the risk of upper gastrointestinal bleeding. Am J Med 1983; 74: 351: 104-7
79-82 75. Rubin PC, Rutherford JM. Drug therapy in pregnant and breast-
55. Graham GG, Graham RI, Day RO. Comparative analgesia, feeding women. In: Carruthers SG, Hoffman BB, Melmon KL,
cardiovascular and renal effects of celecoxib, rofecoxib and et al., editors. Clinical pharmacology. 4th ed. New York:
acetaminophen (paracetamol). Curr Pharm Des 2002; 8: McGraw-Hill, 2000: 1117-42
1063-75 76. Lim H, Paria BC, Das SK, et al. Multiple female reproductive
56. Curhan GC, Willett WC, Rosner B, et al. Frequency of analgesic failures in cyclooxygenase 2-deficient mice. Cell 1997; 91:
use and risk of hypertension in younger women. Arch Intern 197-208
Med 2002; 162: 2204-8 77. Ostensen M. Non-steroidal anti-inflammatory drugs during
57. Barrett BJ. Acetaminophen and adverse chronic renal outcomes: pregnancy. Scand J Rheumatol Suppl 1998; 107: 128-32
an appraisal of the epidemiological evidence. Am J Kidney Dis
78. Stika CS, Gross GA, Leguizamon G, et al. A prospective
1996; 28 Suppl. 1: S14-9
randomized safety trial of celecoxib for treatment of preterm
58. Fored CM, Ejerblad E, Lindblad P, et al. Acetaminophen, aspi- labour. Am J Obstet Gynecol 2002; 187: 653-60
rin and chronic renal failure. N Engl J Med 2001; 345: 1801-8
79. Li DK, Liu L, Odouli R. Exposure to non-steroidal anti-inflam-
59. Lages B, Weiss HJ. Inhibition of human platelet function in matory drugs during pregnancy and risk of miscarriage: popu-
vitro and ex vivo by acetaminophen. Thromb Res 1989; 53: lation based cohort study. BMJ 2003; 327: 368-72
603-13
60. Fischereder M, Jaffe JP. Thrombocytopenia following acute
acetaminophen overdose. Am J Hematol 1994; 45: 258-9 Correspondence and offprints: Professor Garry G. Graham,
Department of Clinical Pharmacology, St Vincent’s Hospi-
61. Bougie D, Aster R. Immune thrombocytopenia resulting from
sensitivity to metabolites of naproxen and acetaminophen. tal, Darlinghurst, NSW 2010, Australia.
Blood 2001; 97: 3846-50 E-mail: ggraham@stvincents.com.au

 2005 Adis Data Information BV. All rights reserved. Drug Safety 2005; 28 (3)

View publication stats