Paracetamol (acetaminophen) overdose and hepatotoxicity: mechanism, treatment, prevention measures, and estimates of burden of disease

ABSTRACT

Introduction

Paracetamol is one of the most used medicines worldwide and is the most common important poisoning in high-income countries. In overdose, paracetamol causes dose-dependent hepatotoxicity. Acetylcysteine is an effective antidote, however despite its use hepatotoxicity and many deaths still occur.

Areas covered

This review summarizes paracetamol overdose and toxicity (including mechanisms, risk factors, risk assessment, and treatment). In addition, we summarize the epidemiology of paracetamol overdose worldwide. A literature search on PubMed for poisoning epidemiology and mortality from 1 January 2017 to 26 October 2022 was performed to estimate rates of paracetamol overdose, liver injury, and deaths worldwide.

Expert opinion

Paracetamol is widely available and yet is substantially more toxic than other analgesics available without prescription. Where data were available, we estimate that paracetamol is involved in 6% of poisonings, 56% of severe acute liver injury and acute liver failure, and 7% of drug-induced liver injury. These estimates are limited by lack of available data from many countries, particularly in Asia, South America, and Africa. Harm reduction from paracetamol is possible through better identification of high-risk overdoses, and better treatment regimens. Large overdoses and those involving modified-release paracetamol are high-risk and can be targeted through legislative change.

KEYWORDS:

• Acetaminophen
• acetylcysteine
• hepatotoxicity
• overdose
• poisoning
• poisons information center
• paracetamol

1. Introduction

Paracetamol (acetaminophen) is one of the most widely used medicines worldwide and is readily available without prescription in most countries [1]. It is listed on the World Health Organisation’s (WHO) Essential Medicines List [2]. It is recommended as a first-line treatment for most cases of pain and fever and is safe to use in children as young as one-month old as well as women who are pregnant [3]. It comes in a variety of forms and strengths including oral tablets, capsules, and liquid formulations as well as rectal suppositories.

Compared to other analgesics accessible without a prescription, paracetamol has a relatively narrow therapeutic index. Toxicity can occur following dosing errors, accidental exploratory ingestions in children, and deliberate self-harm overdose. Paracetamol can cause severe hepatotoxicity with as little as 10 g (or 200 mg/kg for patients under 50 kg) in an acute overdose. Repeated supratherapeutic ingestion can cause toxicity in doses only slightly above the maximum daily therapeutic dose [4]. There is an effective antidote to paracetamol (N-acetylcysteine or NAC); however, despite this, paracetamol toxicity is still the leading cause of acute liver failure (ALF) in most high-income countries [5–10]. Severe cases may require liver transplantation or result in death [9,11].

This review explores recent estimates in rates of overdose and hepatotoxicity from paracetamol, including mechanism, diagnosis, and treatments. We explore global patterns of paracetamol poisoning epidemiology. For the purposes of this review, we define poisoning as contact, whether intentional or accidental, with a substance which may cause harm.

1.1. Literature search

We performed two separate literature searches for this review.

We attempted to analyze available data to determine the number of paracetamol poisoning exposures worldwide (Section 5). When data on paracetamol specifically were not available, we used data on analgesics. To do this we used three methods (Figure 1). The first was a PubMed search for overall poisoning epidemiology literature limited to the last 5 years and manually screened for relevant studies containing data on analgesics and/or paracetamol (see Supplementary Methods). The second was a search of available annual reports from poisons information center websites using the 2021 Poisons Centre Directory available from the WHO. Lastly, using the WHO’s 2021 directory we emailed each PIC for their most recent annual report or data on paracetamol exposures. From this, we were able to include 88 relevant publications, 17 annual reports or statistical information from websites and data from 10 e-mail responses (Figure 1, Supplementary Table S1). For additional details on methodology, see supplementary methods.

Figure 1. Flow chart showing identification and screening of data sources.

The rest of the review covers a wide range of topics on clinical toxicology of paracetamol. For this, we conducted searches on PubMed and Embase from inception to December 2022. We used keyword searches and focused on specific points, including risk assessment, risk factors, management, and legislative change/restrictions. We also looked for systematic reviews relating to paracetamol and conducted citation chaining.

2. Discovery and use

Paracetamol is an aniline derivative. Acetanilide was the first aniline derivative to be used as an antipyretic and analgesic in 1886 [1,12]. However, it was found to have frequent toxic effects (particularly methaemoglobinemia) [13] and so exploration of further aniline derivatives commenced [1]. This led to the discovery of phenacetin and paracetamol in the late 1800s [14,15]. Phenacetin was deemed safer and was used frequently for around 50 years [1,14]. However, in the late 1940s it was discovered that paracetamol was the less toxic major metabolite of phenacetin and acetanilide, and that it did not cause methemoglobinemia [1,16]. This combined with a low risk of nephrotoxicity resulted in paracetamol becoming available commercially in the 1950s [1,14]. Further safety concerns about aspirin in the 1970s led to paracetamol’s current first-line use as an antipyretic and analgesic [1,14].

The mechanism of action of paracetamol is still not entirely understood. Antipyretic effects are thought to be due to the inhibition of prostaglandin production [17]. However, it does not display any anti-inflammatory effects (cf. non-steroidal anti-inflammatory drugs (NSAIDs)), suggesting that it only acts centrally rather than peripherally [17].

Analgesic effects may also be due to interference with descending serotonergic pain pathways through their activation [17]. As such these analgesic effects can be inhibited by serotonin antagonists [17,18]. Paracetamol is generally regarded as a first-line analgesic and is preferred over NSAIDs in certain patient groups. Despite this, a recent systematic review found there was sufficient evidence for the use of paracetamol in only 4 of 44 painful conditions and evidence it was not effective for some common indications [19].

In therapeutic use, paracetamol is rapidly absorbed reaching therapeutic levels after 30 min and peak plasma concentrations within 2 h [20]. Immediate release paracetamol can be taken every 4–6 h and modified/extended release (MR/ER) every 8 h to maintain therapeutic concentrations [3]. Most paracetamol is metabolized by glucuronidation and sulfation and these metabolites are then renally excreted (Figure 2) [1,10,21]. The remaining small amounts are converted into N-acetyl-p-benzoquinoneimine (NAPQI), a toxic metabolite, or excreted unchanged in the urine [1,10,21].

Figure 2. Toxic mechanism of paracetamol and mechanism of action of acetylcysteine. Paracetamol is primarily detoxified by glucuronide and sulfate conjugates which are then excreted in the urine. A small percentage is metabolized by CYP2E1 to the reactive intermediate NAPQI. Under normal conditions, NAPQI can be detoxified by reaction with glutathione to form cysteine and mercaptopuric acid conjugates. If glutathione is depleted (e.g. in paracetamol overdose), NAPQI binds to cell macromolecules causing hepatocyte cell death. The antidote acetylcysteine replenishes cysteine, which is the rate-limiting factor for glutathione synthesis.

3. Toxicity

3.1. Mechanism and manifestations of toxicity

The main toxic effect of paracetamol is hepatoxicity. Paracetamol is a ‘pro-poison’ that exerts its toxic effect through the toxic reactive metabolite, NAPQI. This metabolite is formed by cytochrome P−450 (CYP) enzymes, primarily CYP2E1 and CYP3A4. NAPQI is formed in small amounts in therapeutic doses where it is readily detoxified by conjugation with glutathione [1,10,21]. In overdose, there may be insufficient glutathione to detoxify NAPQI, causing it to bind to cellular proteins (adduct formation) (Figure 2) [10,21]. NAPQI primarily binds to cysteine residues but can potentially also damage proteins at methionine, tryptophan, and tyrosine residues [22]. The mitochondria are a key target for NAPQI adduct formation. Formation of reactive oxygen species causes oxidative stress and leads to activation of c-jun N-terminal kinase (JNK) [23]. The JNK enzyme translocates to the mitochondria, leading to mitochondrial dysfunction, cessation of ATP formation, and mitochondrial membrane rupture [23]. This leads to cellular necrosis. The role of mitochondria in paracetamol hepatotoxicity has been reviewed extensively [24–27]. Severe liver injury leads to loss of hepatic synthetic function and coagulopathy and hypoglycemia. Loss of hepatic metabolic functions leads to encephalopathy and lactic acidosis [20,28]. The clinical manifestations of hepatotoxicity are delayed, with peak serum transaminase levels occurring two to three days after the overdose [10,28]. Approximately 12–13% of acute overdoses result in hepatotoxicity even with treatment [29], with 2–5% progressing to liver failure [30] and 0.2–0.5% resulting in death [30].

Acute kidney injury can also occur, even in the absence of liver failure, and may be delayed [31,32]. Nephrotoxicity may be direct due to tubular necrosis [31–34] from renal NAPQI production or indicate hepatorenal syndrome [32,33].

Paracetamol is also a direct mitochondrial toxin and at very high concentrations can result in central nervous system (CNS) depression. Coma may occur in the absence of hepatotoxicity or other drugs causing CNS depression and may lead to delayed diagnosis and treatment of paracetamol overdose [35–37].

3.2. Risk factors for toxicity

Various factors can increase the risk for toxicity, predominantly by affecting paracetamol’s metabolic pathways.

3.2.1. Massive overdoses

Ingestion of ‘massive’ overdoses (>30 g) increases production of the reactive metabolite NAPQI (Figure 2). This will be exacerbated when the major metabolic pathways become saturated [10,20]. Sulfation saturates at lower concentrations than glucuronidation [38], which saturates with massive overdose. The risk of hepatotoxicity is between 13%−24% in patients who ingest a massive overdose [39–41]. Hepatotoxicity may occur despite early acetylcysteine treatment within 8 h of the ingestion [39–41].

3.2.2. Delayed presentation

Symptoms of hepatotoxicity may be delayed in onset for up to 48 h after ingestion [28]. They include nausea, vomiting and abdominal pain, or tenderness, and may prompt an individual to seek care [10]. Acetylcysteine therapy should be initiated within 8 h of ingestion to be most effective (described below, Section 4.2) [42]. Therefore, late presenters are at greater risk and often already have severe hepatotoxicity [20,29].

3.2.3. Enzyme inducers

Medications which induce the activity of the cytochrome P450 (CYP) enzyme CYP2E1, such as carbamazepine or isoniazid, can theoretically increase the production of NAPQI and contribute to toxicity [43].

3.2.4. Glutathione depletion

Detoxification of NAPQI is dependent on glutathione levels. Individuals who are fasting, malnourished, or anorexic are at increased risk for toxicity due to lower glutathione levels [44]. Hepatotoxicity may occur despite ingestions or concentrations below the usual toxic thresholds [45].

3.2.5. Alcohol consumption

Chronic alcohol consumption may be associated with an increased risk of hepatotoxicity following paracetamol overdose [45]. This may be due to the CYP2E1 induction [43,46,47] or poor nutrition and lower glutathione stores [46]. Caution is given against regular paracetamol use in this setting. Chronic alcohol users are a population more likely to take overdoses, and who often present late, which could be responsible for poorer outcomes [48]. People who chronically consume large quantities of alcohol have also been reported to take larger paracetamol overdoses [49]. Thus, this association likely reflects both behavioral factors and metabolic factors.

In contrast, acute ingestion of alcohol at the time of paracetamol overdose is protective against the development of toxicity [50]. Ethanol is also a substrate for the CYP enzymes responsible for NAPQI generation and competes with paracetamol for CYP2E1, reducing the amount of NAPQI being produced [43].

3.2.6. Ingestion of modified release (MR) and extended release (ER) paracetamol

Sustained release formulations of paracetamol include MR and ER paracetamol. MR paracetamol is available in several countries, including Australia and New Zealand. MR paracetamol contains 665 mg paracetamol, 31% immediate release, and 69% slow release [51,52]. ER paracetamol is available on the US market, and contains 650 mg paracetamol, 50% immediate release, and 50% slow release [53]. Both MR and ER formulations are designed to provide pain relief for up to 8 h.

Overdose with MR paracetamol has been associated with very prolonged absorption and multiple peaks [54] in paracetamol concentrations may occur after overdose and patients often require prolonged treatment due to continuously elevated concentrations [51,54,55]. There may be formation of pharmacobezoars [51,56] which makes absorption even more erratic and this may be seen in massive overdoses of immediate release paracetamol as well [39,57]. There is a higher risk of hepatotoxicity even with early treatment, likely due to the generally higher doses and a mismatch between the timing of acetylcysteine and high paracetamol concentrations.

A retrospective review found that in Australia there was an increasing rate of paracetamol overdose with the MR formulation and that they often involved a larger number of tablets [58]. In addition, deaths from this formulation accounted for a third of the total cases recorded between 2009 and 2017 [58].

For ER paracetamol, early data suggested similar risk between immediate release and ER paracetamol overdose [59]. However, there have since been reports of delayed peaks and delayed nomogram line crossing (where the initial level was below the treatment line, and subsequent levels are above the line) [55]. While it is generally agreed that the Rumack-Matthew nomogram (described below, Section 3.3.1) cannot be safely used for risk assessment in MR and ER overdose [60], it is unclear whether the documented risks for MR paracetamol extend to the ER formulation. MR paracetamol has a longer T_max than ER paracetamol [60]. In addition, MR and ER products are formulated differently. MR tablets use a hydroxypropylmethylcellulose coating [53], and ER tablets are a bilayered tablet where one side contains immediate release and the other slow release paracetamol (in a matrix formulation) [59]. More studies are needed to ascertain the risks of ER paracetamol compared to immediate release and MR formulations. In the absence of these, it is prudent to assume that the risk is higher for all sustained release paracetamol preparations.

3.2.7. Repeated supratherapeutic ingestion and staggered overdoses

Repeated supratherapeutic ingestions occur when patients exceed the recommended doses of paracetamol over a prolonged period for therapeutic purposes. Dental pain appears to be associated with repeated supratherapeutic ingestions, when compared to other types of pain [61]. Repeated supratherapeutic ingestion carries an increased risk of hepatotoxicity when compared to acute single overdose. In a single-center study from the US, hepatotoxicity occurred in 52% of patients with accidental (repeated supratherapeutic) overdose, compared to 14% of those with deliberate overdose [62]. This was despite the accidental group ingesting smaller overdoses [62]. The accidental cases presented later and were more likely to have chronic alcohol use disorder, likely contributing to the increased morbidity [62]. Other studies have compared repeated supratherapeutic and deliberate paracetamol overdose and similarly found therapeutic misadventure to be a risk factor for hepatotoxicity and death [63,64]. Those that develop significant hepatotoxicity typically have elevated aminotransferases on presentation [65,66].

Repeated overdoses can also occur with self-harm intent (often termed ‘staggered’ overdoses). Similar to those with therapeutic intent, these exposures have increased risk of hepatotoxicity when compared to acute self-harm overdoses [67]. This is likely due to delayed presentation [67].

3.2.8. Other risk factors

While the above factors are the best-established risk factors for toxicity following paracetamol overdose, other potential risk factors are emerging. A large retrospective study in the US found significantly increased risk of acute liver injury in people infected with hepatitis C virus, and people with nonalcoholic fatty liver disease (NAFLD) [44]. Hepatitis C virus infection was also associated with higher risk of progression to severe liver disease and mortality [44].

NAFLD has also been identified as a risk factor for hepatic injury following paracetamol overdose [44]. The increased risk from NAFLD has been estimated as similar to that from alcoholic liver disease [68]. This risk is likely related to NAFLD, rather than obesity itself, as some studies have shown no increased risk of paracetamol hepatotoxicity in obese individuals compared to non-obese individuals [69,70]. It has been postulated that the risk of paracetamol hepatotoxicity in an obese individual will depend on the presence of NAFLD and the balance of protective factors (e.g. higher glucuronidation, increased volume of distribution, slower absorption, and reduced CYP3A4 activity) and negative factors (increased CYP2E1 activity, lower glutathione stores) [71].

Genetic polymorphisms in enzymes responsible for paracetamol metabolism have been identified as having the potential to increase risk of hepatotoxicity. Heruth et al [72] have reviewed 147 single nucleotide polymorphisms potentially associated with paracetamol hepatotoxicity. Further studies are needed to ascertain the clinical relevance of these polymorphisms [72]. It is possible that genetic variability could explain elevations in alanine aminotransferase (ALT) in healthy adults administered paracetamol at the maximum therapeutic dose of 4 g per day, and the rare cases of acute liver injury at therapeutic doses [72]. In a genetic study of subjects taking therapeutic paracetamol, 20 genes implicated in paracetamol metabolism and drug-induced liver injury (DILI) development were assayed [73]. The only gene associated with maximum ALT was SULT1E1, which produces the sulfotransferase family 1E member 1 enzyme [73].

3.3. Risk assessment of paracetamol overdose

Biomarkers help determine risk of toxicity, need for treatment, and treatment effectiveness. The usual tests used to guide the management of paracetamol overdose are: serum paracetamol concentration, liver function tests (particularly alanine and aspartate aminotransferases, ALT & AST), and international normalized ratio (INR).

3.3.1. Paracetamol concentrations and the Rumack-Matthew nomogram

Paracetamol concentration is plotted against time from ingestion on the Rumack-Matthew nomogram [74]. This nomogram is used to help in estimating the risk of liver toxicity, and thus whether treatment with acetylcysteine is indicated (see Section 4.2) [75]. Concentrations lying above the line indicate that patients require acetylcysteine and concentrations more than double the line indicate an even greater risk [74]. Some guidelines indicate an increased dose of acetylcysteine in these cases [74]. The Rumack-Matthew nomogram is not useful to predict risk for late presenters (>24 h) or repeated overdoses [74,75].

While the Rumack-Matthew nomogram is used in most countries, the United Kingdom (UK) now uses a nomogram with a lower line (half of the original treatment threshold) [76–78]. This new nomogram was introduced in 2012 due to an isolated fatal case when the paracetamol concentration was below the nomogram line and no treatment was given [77,78]. The lower nomogram line has greatly increased the number of patients receiving acetylcysteine, increasing healthcare costs [77,79]. These patients may be unnecessarily exposed to acetylcysteine and potential adverse reactions [76,77]. Further, various studies have shown that patients with lower paracetamol levels have a greater rate of experiencing adverse effects when treated with NAC [77,80].

3.3.2. Alanine and aspartate aminotransferase (ALT & AST)

ALT and AST are hepatic enzymes that are released into the blood following hepatic cellular damage and are used to diagnose or exclude acute liver injury [42,76]. Therapeutic doses of paracetamol may cause small rises in hepatic enzymes with prolonged treatment [81]. However, aminotransferases return to normal ranges with continued treatment. After an acute overdose ALT/AST may increase after 8 h [42], however they can remain within the normal range for up to 24 h before hepatotoxicity becomes evident [82].

The rate of ALT/AST rise appears important, with initial aminotransferase levels the highest in the patients who developed the most severe toxicity [83]. Between presentation and aminotransferase peak, ALT and AST typically rise in a 1:1 ratio [84]. ALT is generally slower to fall than AST, and an AST:ALT ratio < 0.4 indicates that aminotransferases have peaked and are falling [84]. If the AST:ALT ratio is > 2 at presentation, this may herald worse outcomes, and also has been associated with chronic excessive ethanol use [84]. Increased severity of hepatotoxicity is also associated with a slower fall in AST and in the AST:ALT ratio [84].

Paracetamol-induced liver injury causes marked elevations in aminotransferases, in comparison to most other causes of liver injury. Paracetamol-induced hepatotoxicity is generally defined as either AST or ALT > 1000 IU/L. In patients with severe toxicity, aminotransferase concentrations > 10,000 IU/L are possible; however, once hepatotoxicity is diagnosed, prognosis is better indicated by other markers e.g. INR, pH, and lactate (see Section 3.3.5).

ALT/AST is often measured at presentation and periodically to indicate the need for continued therapy. It is worth noting that ALT/AST is not specific for liver injury and may also lack sensitivity in early liver injury [42,85,86].

3.3.3. The paracetamol-aminotransferase multiplication product

The multiplication product of ALT and paracetamol concentration has been proposed as a risk-stratification tool for predicting hepatotoxicity following paracetamol overdose [87]. One benefit is that this method can be used when exact time of ingestion is unknown, and for staggered and supratherapeutic overdoses. The diagnostic accuracy of this product has been confirmed, with an initial multiplication product of > 10,000 mg/L x IU/L being strongly associated with hepatotoxicity [88]. Patients with a product of <1500 mg/L x IU/L were very unlikely to develop hepatotoxicity [88]. This tool has also been shown to be accurate for MR paracetamol overdose [89]. A recent systematic review has concluded that the paracetamol-aminotransferase multiplication product can complement the Rumack-Matthew nomogram, especially for delayed presentations and staggered ingestions [90].

3.3.4. International normalized ratio (INR)

INR is used to determine the severity of hepatotoxicity as the liver is responsible for the production of many clotting factors [42]. However, paracetamol overdose and acetylcysteine can both directly cause a mildly prolonged INR and thus the test is only useful in established hepatotoxicity [91]. INR levels are then used to indicate severity, prognosis, and the duration of treatment with acetylcysteine [74].

3.3.5. Liver unit referral criteria

These investigations guide treatment with acetylcysteine and stratify risk. Guidelines vary internationally, however generally ALT >1000 U/L is considered an indicator of severe liver injury, and INR > 3 indicates acute liver failure (ALF) and warrants liver unit referral [74]. Other criteria that warrant referral to a specialized liver unit include: oliguria or creatinine > 200 micromol/L, acidosis (pH < 7.3), or lactate >3 mmol/L, hypoglycemia, thrombocytopenia, hypotension, and encephalopathy (see Section 4.7) [74].

3.3.6. Novel biomarkers

There are many other potential biomarkers to predict the development of liver toxicity and guide treatment [42,75,76,86]. Ideal biomarkers would detect liver damage earlier than ALT and INR and also predict severity and prognosis. Promising candidates include mitochondrial damage and necrosis markers and paracetamol protein adducts which measure the toxic metabolite NAPQI bound to cellular proteins [42,75,76,82]. Paracetamol-protein adducts have been shown to be marginally superior to traditional biomarkers at predicting acute liver injury and can diagnose paracetamol-induced hepatotoxicity in the absence of a history of ingestion [38,82,92,93].

4. Treatment of paracetamol overdose

4.1. Charcoal

Activated charcoal binds paracetamol and thus may reduce the amount of paracetamol absorbed from the gastrointestinal tract [94]. It may be useful in acute paracetamol overdoses that present to primary care facilities within 2 h of ingestion (or 4 h for ‘massive’ overdoses or exposures to MR/ER preparations) [74]. However, it is not useful in patients who ingest liquid formulations due to the rapid absorption [74]. It is also not recommended after repeated supratherapeutic ingestions [74]. Charcoal may reduce peak paracetamol concentrations, the need for NAC and decrease the risk of hepatotoxicity [39,95].

4.2. N-acetylcysteine (NAC)

NAC is the mainstay of paracetamol overdose management. NAC restores glutathione by replenishing cysteine, the rate-limiting factor for glutathione synthesis [96,97]. This allows for the detoxification of NAPQI and prevents hepatotoxicity in most cases [98]. The risk of developing hepatotoxicity is substantially reduced when NAC is given within 8 h of ingestion [21].

Before the availability of NAC treatment, morbidity and mortality following paracetamol overdose was considerable; many developed acute liver injury and 3–5% died [99,100]. Since acetylcysteine has become standard treatment, deaths are rare, less than 1% in most cohorts [94].

Various regimens for NAC have been used. NAC was originally delivered orally until the development of an intravenous formulation [101]. Both routes are effective, although the intravenous formulation is now more commonly used, and has advantages when there are factors that may affect oral absorption including vomiting, gastrointestinal conditions or evidence of hepatic failure [29]. Several intravenous regimens have been the subject of clinical trials and the ‘3-bag regimen’ was most common for many years [74,94,101]. The use of simplified ‘2-bag’ regimens reduces adverse drug reactions but has comparable efficacy [74,102–104]. The simplest modification is a 200 mg/kg bag delivered over 4 h, followed by a 100 mg/kg bag delivered over 16 h [74]. The two or three bag regimens can also be combined into one infusion bag given with variable rates, which may reduce the risk of gaps in treatment [105].

Abbreviated NAC regimens are sometimes used in the treatment of repeated supratherapeutic ingestion. In Australia, an 8-h NAC regimen is used in repeated supratherapeutic ingestions where ALT remains static and paracetamol concentrations remain low [74]. A retrospective study indicated that this is likely safe (there were no cases of re-presentation with hepatotoxicity) and reduces hospital length of stay [106]. In the US, the Rocky Mountain Poison and Drug Center protocol treats repeated supratherapeutic ingestion patients with 12 h of NAC, with treatment only being continued if ALT/AST is rising or the paracetamol concentration is elevated [107].

There is increasing recognition that dosing NAC based on body weight may not be sufficient to prevent injury in patients taking high doses of paracetamol. This has led to the practice of increased NAC dosing in high-risk patients. For example, doubling the dose of NAC given in the second/third bag when paracetamol concentrations are more than double the nomogram line reduces the risk of hepatotoxicity [39,74]. Optimal NAC dosing regimens and future directions have been recently reviewed [108].

Biomarkers including hepatic enzymes, INR, and paracetamol concentrations are used to guide the duration of NAC [74]. Extended NAC infusions should be given in high-risk patients (e.g. patients with persistently elevated paracetamol concentrations or established hepatotoxicity) [74].

4.3. Methionine

Prior to the active use of acetylcysteine, other treatments such as methionine were used [94,99,109]. Similar to NAC, methionine is an antidote used to increase glutathione and treat or prevent hepatotoxicity [109]. Although primarily administered orally, it has been administered intravenously [99,109,110]. Methionine may cause gastrointestinal, neurological, or cardiovascular side effects which limit its acceptability [109,110] and so it is no longer routinely recommended. Although initially regarded by the WHO as having a similar efficacy and safety profile as acetylcysteine [94], methionine is no longer on the most recent WHO Essential Medicines List [2].

4.4. Cimetidine

Cimetidine is a histamine−2 receptor antagonist and an inhibitor of CYP2E1 [111]. Based on this inhibition, it has been proposed as a potential treatment for paracetamol overdose [111]. There are many animal studies showing that cimetidine reduced hepatotoxicity from paracetamol, summarized by Mullins et al [111]. However, disappointingly, clinical effectiveness in humans has not been demonstrated, albeit these have been small studies with few high-risk patients [111].

4.5. Fomepizole

Fomepizole is best known as a treatment for toxic alcohol poisoning, where it inhibits alcohol dehydrogenase [112]. Fomepizole has recently been proposed as a treatment to prevent hepatotoxicity following paracetamol overdose. The mechanism is unrelated to alcohol dehydrogenase but is thought to be a result of two functions [112]. Fomepizole inhibits CYP2E1, reducing conversion of paracetamol into NAPQI [112]. Secondly, it can also prevent the activation of the JNK enzyme and resultant mitochondrial dysfunction [112,113].

The evidence for use of fomepizole in this context is primarily from animal models [113] and small case series [114,115]. Interpretation of results from observational case series has been controversial [112,116,117]. Some have advocated for the use of fomepizole as an adjunct to NAC in massive overdoses [114]. Others have called for the need for large comparative studies to generate evidence before fomepizole is considered a recommended treatment [116]. If the promising early results are replicated in robust clinical trials, fomepizole could fill an un-met need in this space, particularly for patients where NAC is less effective (massive overdose, delayed presentation).

4.6. Calmangafodipir

Manganese superoxide dismutase is a mitochondrial protein that scavenges free radicals and can prevent mitochondrial injury. Toxic doses of paracetamol inhibit manganese superoxide dismutase in animal studies [118]. Restoring or mimicking manganese superoxide dismutase activity could be used as a treatment for paracetamol overdose. Calmangafodipir is a superoxide dismutase mimetic [119]. It has been successfully used in Phase 2 trials for chemotherapy induced peripheral neuropathy mediated by oxidative stress. In a small study where calmangafodipir was used as an adjunct to NAC in paracetamol overdose, no safety issues were identified [119]. That trial also found that calmangafodipir may reduce liver injury biomarkers, however the sample size and design was not powered to study efficacy, and most patients enrolled were low-risk for hepatotoxicity [119]. Further studies designed to assess efficacy are needed. Fomepizole, calmangafodipir, and other potential therapeutic targets for paracetamol hepatotoxicity have been recently reviewed [120].

4.7. Liver transplant

In patients with severe liver failure and a grave prognosis, orthotopic liver transplant is considered [121]. Kings College Criteria are widely used to predict patients who are likely to die without a liver transplant [121,122]. There is some controversy surrounding the benefits of liver transplantation following paracetamol overdose, with a systematic review suggesting that transplant may confer little long-term survival benefit [121,123].

4.8. Lack of global treatment guidelines

Clinical practice guidelines allow for consolidation of available materials to guide areas such as diagnosis, treatment recommendations and monitoring. They are generally easy to navigate and provide clear instructions on the course of management for the suspected condition. Their use is invaluable in a variety of conditions with potentially severe outcomes, including paracetamol poisoning. However, other than the development of the Australian and New Zealand guidelines, those providing guidance on the management of this condition are not readily available or freely accessible [74]. Many countries may have established national or regional guidelines, although it is possible that they are only accessible through subscription-based online resources or via contact with a Poisons Information Centre (PIC). However, the development of consistent global guidelines is difficult as the treatment of paracetamol poisoning is not a ‘one size fits all’ approach. Various countries have adopted alternate NAC regimens or even lower cut-off limits to start treatment based on the Rumack-Matthew nomogram [101]. Denmark even forgoes the use of this nomogram and treats all patients with a full course of NAC regardless of paracetamol concentrations [80].

5. Epidemiology of paracetamol overdose

5.1. Methodology for estimating the impact of paracetamol overdose globally

The search strategy conducted for this section of the review is described above (Section 1.1) and in further detail in Supplementary Methods.

The data is presented as yearly counts per 100,000 population where a catchment population was able to be determined and/or as a percentage of all poisoning exposures. Where possible, we calculated case fatality rates by dividing the number of cause-specific poisoning fatalities by the number of exposures.

Weighted averages were calculated using meta-analysis to account for heterogeneity between the data sources. Meta-analysis was performed in R 4.0.4. For full details, including definitions of analgesics, DILI, and ALF/ALI, see Supplementary Methods.

5.2. Data availability by region

There were a total of 115 data sources containing information about analgesic and/or paracetamol poisoning. Countries in Europe, North America, and Asia were the greatest sources of contribution (Supplementary Figure S1). The available data is not proportional to population [124], and many continents provided little data. Some publications included data from multiple countries however these were often countries within the same geographical area. Within countries, some data sets contain information from certain regions only. In addition, data sources available differed by country. As such in the figures, the data points for each country have been color coded according to the type of data source: poisons information centers, a single hospital/treatment center and multiple hospitals/treatment centers or large administrative data sets.

5.3. Overall and intentional exposures

5.3.1. Rates as a percentage

Where data were available, we found a weighted mean of 9.7% (range 0.7–40.2%) of poisoning cases were attributable to analgesics (Supplementary Figure S2) [125–157]. Denmark had the highest proportion of poisonings attributable to analgesics at 40.2%, noting that this was a proportion of drug poisonings only rather than all poisonings [133]. The lowest rate was reported by a German Poisons Information Centre [151].

Paracetamol accounted for an average of 6.2% (range 0.5–34.9%) of poisonings where data were available (Figure 3) [125–131,133,135,136,138,140,142–144,148,150,156–183] (including additional data from e-mail responses, Supplementary Table S1). It accounted for over a quarter of all drug poisonings in Denmark [133] and of all poisonings in a cohort of > 17 000 patients in Australia [159]. Croatia had very few paracetamol poisonings (<1%) [162].

Figure 3. Percentage of poisonings relating to paracetamol. Dark blue points represent data from a single hospital/treatment center, light blue points represent data from multiple hospitals/treatment centers or large administrative databases, while orange points represent poisons information center data. Dotted line shows weighted mean across sources 6.2%.

*Refers to data sources using the ToxIC case registry which is based on data predominantly from the U.S.A but may also include data from Canada, Australia, Israel, Saudi Arabia, and Thailand. One data source in Belgium also contains data from Luxembourg. One source in Brazil contains data from a Poisons Information Centre combined with data from multiple hospitals/treatment centers. Data sources in Chile and Denmark reported poisonings from drugs/pharmaceuticals only. One source from Portugal (3.8%) reported accidental/unintentional poisonings only. One data source in the U.S.A (16.5%), and the Ethiopian source reported poisonings from women only.

Intensive care unit (ICU) poisonings represent a severe subset of poisonings and were not pooled with overall rates. Analgesic poisonings account for 8.0% of ICU poisonings in France [184] and 12.4% in Germany [185]. An Iranian source reported that paracetamol accounted for only 0.7% of ICU poisonings [186].

For sources reporting intentional exposures only, analgesic poisonings accounted for 16.3% (range 6.2–29.2%) (Supplementary Figure S3) [135,148,187–192].

Intentional paracetamol poisonings accounted for 11.8% (range 1.8–37.3%) of all intentional poisonings, highest in New Zealand and lowest in Croatia (Figure 4) [125–128,144,148,157,162,193,194] (including additional data from e-mail responses, Supplementary Table S1).

Figure 4. Percentage of poisonings caused by intentional ingestion of paracetamol. Dark blue points represent data from a single hospital/treatment center, light blue points represent data from multiple hospitals/treatment centers or large administrative databases, while orange points represent poisons information center data. Dotted line shows weighted mean across all sources: 11.8%.

*Refers to data sources using the ToxIC case registry which is based on data predominantly from the U.S.A but may also include data from Canada, Australia, Israel, Saudi Arabia, and Thailand.

5.3.2. Rates per 100,000 population per year

We found a weighted mean of 11.5 (range 0.1–88.9) analgesic poisonings per 100,000 population per year, lowest in China and Botswana (Supplementary Figure S4) [125–128,130–134,136,139,145–157]. On average intentional ingestions of analgesics contributed to 4.9 (range 0.4–26.0) poisonings per 100,000 population per year, with data only available for four countries (Supplementary Figure S5) [148,188,192,195].

Paracetamol was involved in an average of 7.4 (range 0.1–63.6) poisonings per 100,000 population per year (see Figure 5) [58,125–128,130,131,133,136,148,150,156–160,162,163,166,168,169,171,172,174,176–182,196–200] (including additional data from e-mail responses, Supplementary Table S1). The lowest rate was reported in a data source reporting on females only, in Ethiopia [163]. Australia had the highest rate reported at 63.6 per 100,000 population per year [182]. Focusing on intentional exposures, there were 5.4 (range 0.2–31.9) intentional paracetamol poisonings per 100,000 population per year (Figure 6) [58,125–128,148,157,162,194,196,197,200] (including additional data from e-mail responses, Supplementary Table S1).

Figure 5. Paracetamol poisonings per 100,000 population per year. Dark blue points represent data from a single hospital/treatment center, light blue points represent data from multiple hospitals/treatment centers or large administrative databases, while orange points represent poisons information center data. Dotted line shows weighted mean across all sources: 7.4/100,000.

One data source in Belgium also contains data from Luxembourg. One data source in Brazil contains data from a Poisons Information Centre combined with data from multiple hospitals/treatment centers. Data sources in Chile and Denmark reported poisonings from drugs/pharmaceuticals only. One source from Portugal (0.2/100,000) reported accidental/unintentional poisonings only. The data source from Ethiopia only included women in their report.

Figure 6. Intentional paracetamol poisonings per 100,000 population per year. Dark blue points represent data from a single hospital/treatment center, light blue points represent data from multiple hospitals/treatment centers or large administrative databases, while orange points represent poisons information center data. Dotted line shows weighted mean across all sources: 5.4/100,000.

5.4. Rates of hepatotoxicity and liver failure

Overall, paracetamol accounted for a weighted mean of 55.7% (range 50.0–77.8%) of severe acute liver injury (ALI) and/or ALF [201–208] and 7.3% (range 1.2–29.2%) of drug-induced liver injury (DILI) [209–219] (see Figure 7). The country with the highest percentage of DILI cases caused by paracetamol was the U.S.A at almost 30% [217] followed by Pakistan at 16% [215]. For sources examining causes of ALI/ALF, Scotland and Australia had the highest number of cases relating to paracetamol at approximately 78% [204] and 63% [201], respectively.

Figure 7. Proportion of patients who develop ALI/ALF or DILI after paracetamol use. Dark blue points represent data from a single hospital/treatment center and light blue points represent data from multiple hospitals/treatment centers or large administrative databases. Dotted (ALI/ALF) and dashed (DILI) lines show weighted mean across all sources: ALI/ALF = 55.7%, DILI = 7.3%.

5.5. Fatalities

Where data were available, analgesics contributed to 0.1 (range 0.0003–0.7) deaths per 100,000 population per year (Supplementary Figure S6) [125–128,157,195,220–224]. When data sources including opioids were removed, this dropped to 0.04/100,000. Sources reporting on paracetamol only found similar results, with a weighted mean of 0.04 (range 0.0001–0.3) deaths per 100,000 population per year attributable to paracetamol (Figure 8) [58,125–128,148,153,157,160,196,221,222,224–228].

Figure 8. Deaths attributable to paracetamol poisonings per 100,000 population per year. Dark blue points represent data from a single hospital/treatment center, light blue points represent data from multiple hospitals/treatment centers or large administrative databases, orange points represent poisons information center data, purple points represent coronial/forensic databases. Dotted line shows weighted mean across all sources: 0.04/100,000.

Two sources in Brazil are combined data sources – one from a Poisons Information Centre and multiple hospitals/treatment centers and the other from a Poisons Information Center, multiple hospitals/treatment centers, and national datasets. The data source from Switzerland only reported deaths from intentional poisonings. The Australian source used coronial data but each case was reviewed manually by researchers to include only those where the decedent had documented hepatotoxicity.

The case fatality rate for analgesic poisonings averaged 0.7% (range 0.1–2.5%) (Supplementary Figure S7) [125–128,141,157,195]. A German ICU data source reported a case fatality rate of 4.1% but this was not pooled with other data sources due to the bias toward more severe poisonings in the ICU [185]. The case fatality rate for paracetamol poisonings was 0.4% (range 0.1–2.3%) (Supplementary Figure S8) [58,125–128,135,143,144,148,157,160,196,229,230].

5.6. Limitations of available data

There were several limitations impacting the amount and type of data that was collected. By conducting a literature search through PubMed only, some relevant studies may have been missed. Although this data may have been published in annual reports instead, not all countries have a PIC and there is also a clear lack of representation of low- and middle-income countries in the available data. Using the WHO’s poisons center directory, e-mails were sent out to 280 different PICs yet there were only 11 direct responses and 66 e-mails that were undeliverable. Over 50 websites were inaccessible and although there were some PICs with an annual report, many did not report on the data required for this analysis. Many websites were a direct link to the hospital that the PIC worked out of rather than the PIC itself and many did not contain any annual reports. Data presented in languages other than English may have been interpreted incorrectly and as such some data may not be accurate.

Operational models of PICs vary worldwide. Some provide advice to healthcare practitioners only, whereas others also include members of the public. Some act as toxicology services within hospitals, while others provide advice over the phone or via a hybrid model with online databases (such as the UK, New Zealand, and the Netherlands). This could result in variability in rates recorded. Attempts were made to identify catchment populations to allow calculation of population rates, however this was not available for much of the literature resulting in the use of rates as a percentage where available. Some data sources only provided rates without providing raw data or catchment populations [195,196] and as such were excluded from the calculation of the weighted means. Additionally, some countries had data from certain regions only. It is unlikely that particular regions are representative of a whole country, which may reduce the accuracy of these results.

Death data were from a range of data sources including poison centers where outcome data is sought (e.g. U.S.A), single hospitals, and coronial/forensic/national death statistics data. These data sources each have their limitations, for example poisons center and hospital datasets do not capture out-of-hospital deaths, and some patients will be lost to poisons center follow-up. The datasets using coronial/forensic data may over-estimate cases by counting those where paracetamol was detected in multi-drug overdose, but hepatotoxicity did not occur. There was no clear pattern between data source and estimate (Figure 8), indicating that variability may be due to country-specific differences as well as data capture.

Many publications contained exclusions such as age or sex, and we excluded pediatric only data sources as the rates are likely to be very different to whole-of-population rates where intentional overdose predominates. Some sources were limited to drug poisonings only which is likely to result in the variety of estimates seen. Definitions for analgesics were not consistent throughout the literature. We attempted to remove data relating to opioids however sometimes this was not possible as data was already pooled by the original study authors. Poisons center enquiries were not well defined and may include exposures in addition to calls regarding general information as well as recalls for the same exposure, for this reason we only included poisons center data on exposures.

6. Paracetamol availability and its impacts

6.1. Non pharmacy sales, pharmacy sales, and pack size restrictions

Paracetamol is available without a prescription in most countries. It is one of the most common over-the-counter (OTC) analgesics used in suicidal overdoses [231]. A study in Ireland found that paracetamol was used in over half of all overdoses using OTC drugs [232]. The volumes that can be purchased and the availability of non-pharmacy sales differ substantially globally (Table 1). A study of 21 European countries found that over half of the participating countries including France, Italy, and Switzerland did not have paracetamol available for purchase outside of pharmacies and that the rate of paracetamol-related poisoning enquiries was subsequently lower in most of these countries [233]. Countries such as the UK, Denmark, and Ireland allow for sales of paracetamol outside of pharmacies however the pack sizes are limited to 5–8 g [233]. The introduction of paracetamol sales from non-pharmacy outlets in Sweden in 2009 was associated with an increase in the rate of paracetamol poisonings [234], however restrictions were reintroduced in 2015 [233].

Table 1. Restrictions on pack sizes of over-the-counter (OTC) pharmacy and non-pharmacy sales of immediate release paracetamol.

Country	OTC pharmacy sales (g)	Non-pharmacy sales (g)
Australia	50	10
Austria	30	N/A
Belgium	10	N/A
Canada	Unlimited	Unlimited
Croatia	Unlimited	N/A
Czech Republic	Unlimited	6
Denmark	10	5
France	8	N/A
Finland	15	N/A
Germany	10	N/A
Iceland	15	N/A
Ireland	12	6
Italy	15	N/A
Lithuania	Unlimited	N/A
The Netherlands	Unlimited	N/A
New Zealand	50	10
Norway	10	5
Poland	Unlimited	6
Russia	Unlimited	Unlimited
Singapore	Unlimited	Unlimited
Slovakia	Unlimited	N/A
Slovenia	10	N/A
Sweden	10	N/A
Switzerland	8	N/A
UK	16	8
USA	Unlimited	Unlimited

N/A = no sales allowed [30,233].

Over two-thirds of European countries have paracetamol pack size restrictions within pharmacies, ranging from 8 to 30 g [233]. The number of paracetamol-related poisoning enquiries was higher in some of these countries [233].

In the U.S.A and Canada, paracetamol is available outside of pharmacies in very large packs [30]. Paracetamol was involved in almost half of suicides/attempted suicides using OTC analgesics in the U.S.A [235].

In Singapore, paracetamol can be sold outside of pharmacies without restriction [30]. Within Australia, and similarly in New Zealand, products containing up to 20 tablets of 500 mg of immediate-release paracetamol can be sold outside of pharmacies [236,237]. There are no legal limits on the maximum amount (number of packs) of paracetamol that can be purchased outside of pharmacies. In Australia and New Zealand, within pharmacies customers may purchase packs of up to 100 tablets of immediate-release paracetamol without any intervention by a pharmacist [236,237].

6.2. MR/ER Paracetamol

MR paracetamol presents a significant challenge in overdose and is more dangerous than immediate-release paracetamol. It can cause hepatotoxicity despite early treatment with NAC, and treatment with charcoal and increased NAC are not effective in preventing hepatotoxicity even in those who present early.

After a decision made by the European Medicines Agency, countries in the European Union such as the UK (formerly), Denmark, and Ireland do not supply any MR paracetamol, even on prescription [30,238]. On the other hand, countries such as the United States and Singapore have no restrictions on ER paracetamol and allow this formulation to be sold outside of pharmacies as well [30].

In Australia, the MR formulation was originally available in pharmacies without consultation with a pharmacist, similar to the immediate-release paracetamol. In 2020, overdose concerns led to it being up-scheduled such that supply required a pharmacist’s approval (similar changes also occurred in New Zealand) [236, 239].

6.3. Impact of regulatory changes

Worldwide there have been few studies on changes aimed at reducing the occurrence of paracetamol overdose and its complications. Some countries have implemented pack size restrictions and/or moved to restrict access to paracetamol outside of pharmacies.

6.3.1. The United Kingdom

In the UK in 1998 legislation was introduced to reduce the pack sizes of available paracetamol both within and outside of pharmacies [240]. Customers would be allowed to purchase up to 8 g of paracetamol in general stores and up to 16 g within pharmacies but without assistance; any greater amount required discussion with a pharmacist [240]. In the year following the intervention, there was a significant reduction in both the total number of paracetamol overdoses and the number of severe paracetamol overdoses at the Royal Free Hospital [241]. Benzodiazepine overdoses were used as a control, and remained stable [241]. A study that combined Office for National Statistics data with information from six liver units and monitoring systems in general hospitals found consistent results [242]. Suicidal deaths reduced by 22% in the first year, which was sustained for the following two years [242]. Liver unit admissions and liver transplants were reduced by ~ 30% and large overdoses were reduced by 20% [242]. There was a simultaneous increase in ibuprofen overdoses but little-to-no effect on suicides [242].

Some studies of this intervention have had neutral or negative findings. A study using national deaths data and hospital admissions data found declining mortality and hospital admissions from paracetamol between 1993 and 2002 [243]. However, the authors concluded that the contribution of the 1998 intervention was unclear [243]. An interrupted time-series analysis found a decrease in paracetamol poisoning mortality associated with the intervention [244]. Fatal poisoning by some other agents (including paracetamol compounds not affected by the legislation) also decreased [244]. The authors concluded that the reduction seen might have been part of a broader trend [244].

A Scottish study of hospitalizations and deaths from paracetamol found no reduction in paracetamol overdose frequency or mortality following the 1998 intervention [245].

A systematic review examining the effects of this intervention found that there was a significantly reduced amount of paracetamol being purchased and a reduction in the number of admissions to liver units and subsequent transplants [240]. There was also a significant decrease in the total number of deaths caused by paracetamol, albeit with still over 100 deaths per year [246].

Retailer adherence to the legislation has been questioned. A ‘mystery shopper’ study in South London in 2004 tested whether amounts in excess of the 1998 legislation could be purchased from non-pharmacy and pharmacy outlets [247]. Researchers were able to purchase > 32 tablets in four of eight pharmacies, and > 16 tablets in 13 of 16 other retailers [247]. The same study surveyed patients at a London hospital who took paracetamol overdoses. Approximately half of patients ingesting > 16 paracetamol tablets purchased paracetamol for the purpose of taking an overdose [247]. Of these 35 people, 15 went to multiple different stores, whereas 16 were able to purchase multiple packs from one store [247].

Importantly, while maximum pack sizes were mandated, limits to number of packs that could be purchased were not implemented by the UK Medicines and Healthcare Products Regulatory Agency (MHRA). Thus, stores selling more than one pack in the above study were contravening the spirit of the 1998 legislation but were not breaking the law [248]. In 2009, the MHRA introduced guidelines stating ‘no retailer should sell more than two packets of paracetamol (500 mg) in one transaction, retailers should be discouraged from promoting the sale of more than one pack at a time and that it is illegal to sell more than 100 tablets of paracetamol (500 mg) … in one transaction’ [248]. Despite this, a more recent mystery shopper study of non-pharmacy outlets found that 58% of retailers sold more than the recommended amount, and 23% sold more than the legal amount [248].

Taken together, evidence from the UK suggests that changes in paracetamol availability can reduce severe self-poisoning and deaths, but these changes are more modest than initially hoped [249]. Lack of enforcement of guidelines and laws likely reduced the impact of this legislation. Legal limits to the number of packs that can be purchased, and active enforcement should be considered by countries considering similar legislation.

6.3.2. Denmark

Denmark undertook two interventions [250]. The first in 2011 restricted sales to persons aged 18 years and over while the second in 2013 restricted pack sizes to 20 tablets for OTC purchases within pharmacies [250]. After the age restriction, there was a reduction in the number of poisonings in those aged 10–17 years but no change in total non-opioid analgesic poisonings [250]. This restriction may have also played a part in reducing the number of cases of self-harm seen in adolescents in Denmark [251]. The introduction of the pack size restriction saw a decrease in the number of total poisonings and their severity [250].

6.3.3. North America

In 2014, the U.S.A attempted to reduce unintentional overdose, and therefore hepatotoxicity, by reducing the quantity of acetaminophen when combined with opioids to 325 mg per tablet [252]. However, PIC data from before and after the change identified no significant variation to the rates of treatment for poisoning or hepatotoxicity [252]. Canada attempted to reduce the risk of accidental overdose with paracetamol by updating labeling standards on two occasions (2009 and 2016); however, no significant changes occurred [253].

6.3.4. Australia

In response to rising rates of overdose with MR paracetamol in Australia, and the increased risk of toxicity with MR paracetamol, Australia up-scheduled MR paracetamol from ‘Pharmacy Only’ to ‘Pharmacist Only’ in 2020 [239]. This meant that MR paracetamol could only be purchased from behind the pharmacists’ counter, and a pharmacist had to be involved in the sale. Early evaluation of this intervention indicates that it was not associated with a reduction in poisonings with MR paracetamol [254].[253]

7. Conclusions

Paracetamol is one of the most widely used medicines in the world. This is despite a recent systematic review indicating lack of efficacy for many conditions for which it is considered first line. It is widely available without prescription, however access varies greatly internationally. Paracetamol is one of the most used substances in overdose in many parts of the world, and the leading cause of acute liver injury in many countries. Where data were available, we estimate it is used in 6% of poisonings, implicated in 56% of cases of severe acute liver injury and acute liver failure and 7% of drug-induced liver injury. Approximately 0.4% of paracetamol overdose cases are fatal. Acetylcysteine is an effective antidote, and yet hepatotoxicity and death still occur. The three main factors making hepatotoxicity more likely are (i) larger overdoses, (ii) overdose with MR/ER paracetamol, and (iii) delays to treatment. Recent research has focused on changes to acetylcysteine regimens to reduce adverse reactions and improve efficacy for high-risk overdoses, as well as the exploration of novel biomarkers and metabolic treatments. In addition, public health interventions such as pack size restrictions and age limits for purchasing paracetamol have been evaluated and been shown as an effective way at reducing harm from paracetamol overdose.

8. Expert opinion

If paracetamol was discovered today, drug regulators would likely enforce tighter restrictions than those currently in place due to the risk of hepatotoxicity. Despite widespread use and availability, paracetamol has a narrow safety margin compared to other over-the-counter analgesics. For example, the toxic dose of paracetamol is 10 times a therapeutic dose, whereas the toxic dose of ibuprofen is 60 times a therapeutic dose. Worldwide accessibility of paracetamol varies substantially. Broadly, continental Europe has strict rules: many countries do not allow non-pharmacy sales, and sales within pharmacies are typically limited to small packs. The UK reduced available pack sizes in 1998 in response to rising rates of overdose and deaths. In the U.S.A., there appear to be no limits on pack size (1000 tablet packs can be purchased).

The magnitude of the paracetamol problem is unknown in many countries. Despite our best efforts to capture published studies and gray literature, our estimates of the burden of paracetamol poisoning globally may not be internationally representative, and many countries had no data available. This is somewhat surprising, as there is an ICD−10 code for paracetamol overdose, so this information should be collected in most countries. Disappointingly, the global burden of disease studies reports specifically only on carbon monoxide poisoning with all other poisonings combined. Improving global estimates of disease burden from poisoning is vital.

Paracetamol has long been considered the most important common poisoning in high-income countries and is increasingly being used in overdose in low- and middle-income countries [255]. Acetylcysteine is an effective antidote: without it it is estimated that 3–5% of overdoses would be lethal [99,100]. Despite its long-term use, controversies in the management of overdose remain. We found few publicly accessible treatment guidelines, however we know treatment differs substantially globally. This includes the UK lowering of the treatment line in the Rumack-Matthew nomogram, and some countries treating all overdoses, regardless of serum concentrations. Acetylcysteine regimens differ, with new regimens being recently introduced in many countries to reduce adverse effects relating to the large loading dose.

Changing use patterns, including increasing ‘massive’ overdoses and use of modified-release paracetamol present treatment challenges. Recent innovations in N-acetylcysteine dosing regimens, new antidotes, and increased recognition of the role of activated charcoal may improve outcomes for patients. The identification of novel biomarkers has the potential to improve early identification of hepatotoxicity and better target treatments. This is an area of ongoing active research.

Emerging potential antidotes that target mitochondrial effects of NAPQI include fomepizole and calmangafodipir. If early clinical results are confirmed by robustly designed trials, these antidotes could fill a much-needed gap for scenarios where acetylcysteine is less effective. This includes people who present late and with established hepatotoxicity. The clinical toxicology field should learn from the experience with hydroxychloroquine and COVID−19, where early studies (animal experiments, small case series) showed promising results, which were later refuted by a robust platform trial. Given the frequency of paracetamol overdose, well designed, large clinical trials of these antidotes should be possible before they are considered routine care. In 10 years, the face of paracetamol management may look completely different if new biomarkers and new treatments become commonplace.

Continued innovation to improve outcomes has never been more important. Many countries are facing an epidemic of self-poisoning, particularly among adolescent females. This has been observed in the U.S.A, Canada, Australia, Egypt, and Norway, and is likely happening un-documented in many other countries. Paracetamol is the agent taken most frequently by young females, who have higher rates of OTC medicines use in overdose compared to adults (likely due to reduced access to prescription medicines). In 2022, Australia’s national drug regulator launched an enquiry into paracetamol overdoses in young people in response to surging rates.

The drivers behind surging rates of youth self-harm are complex. However, there are simple solutions to reduce harm from paracetamol overdose. Any intervention should be aimed at the most modifiable risk factors for hepatotoxicity: overdose size and overdose with MR/ER paracetamol. Self-harm overdoses tend to be impulsive, many people contemplate their overdose for less than 5 min. People typically take what is available to them in the home, and often consume every tablet they have access to. Reducing the paracetamol ‘tablet burden’ in homes by reducing pack sizes available without prescription is expected to impact overdose size and thus harm. This played out in the UK, with pack size restrictions reducing large overdoses, liver transplants and deaths without harm from methods substitution. However, the UK intervention was not as effective as legislators hoped, and paracetamol poisoning remains a concern. This sub-optimal effect is possibly due to some retailers breaching the spirit of the legislation, as indicated by mystery shopper studies. Other countries considering pack size restrictions should limit the number of packs that can be sold in one transaction. Admittedly, this may be more difficult to legislate and enforce than simple pack size limits.

In addition, reducing the casual use of MR/ER paracetamol would minimize its use in overdoses by vulnerable people. Recent data from Australia indicate that making MR paracetamol only available following consultation with a pharmacist did not curb poisonings. Previous Australian experience with codeine has indicated that more restrictive rescheduling to prescription only had a much stronger effect on poisonings than ‘Pharmacist Only’ rescheduling. The European Medicines Agency suspended marketing of modified-release paracetamol due to an unacceptable safety profile in overdose. The sole therapeutic benefit of MR paracetamol is a slight simplification in dosage regimen (three doses/day rather than four). Restrictions could be considered elsewhere in the world.

The relative lack of studies evaluating paracetamol poisoning interventions indicates that legislative measures are often implemented without evaluation plans. For example, we were unable to find any published data on the impact of the EMA suspending MR paracetamol on paracetamol hepatotoxicity and death. Prospective evaluation of any legislation aimed to reduce harm from poisoning is essential for targeting future interventions and informing responses in other jurisdictions.

It is important to keep this in perspective. Tonnes of paracetamol are consumed worldwide every day, mostly without harm. Paracetamol has a very good safety profile in therapeutic use. It has minimal adverse reactions if taken appropriately, a short list of clinically significant drug interactions, and can be taken by young infants, pregnant women, and people who cannot tolerate NSAIDs. Despite recent efficacy concerns, there are many people for whom paracetamol will remain first-line treatment for pain. Public health measures and continued clinical innovations are needed to ensure an acceptable risk-benefit profile.

Article highlights

• Paracetamol (acetaminophen) is one of the most used medicines worldwide. It is safe if taken at therapeutic doses, however is commonly taken in overdose, where it can cause hepatotoxicity.

• Risk factors for toxicity include: very large overdoses, overdose with modified release paracetamol, delays to treatment, treatment with enzyme inducers, malnutrition, and chronic alcohol use.

• Acetylcysteine is an effective antidote. It replenishes glutathione stores, allowing detoxification of paracetamol’s toxic reactive metabolite. Recently, many changes to acetylcysteine treatment regimens have emerged to better treat high-risk overdoses and minimize adverse reactions. Other emerging potential treatments include fomepizole and calmangafodipir.

• Epidemiological data on paracetamol poisoning are limited or non-existent in many parts of the world. Based on available data, we estimate that paracetamol is involved in 6% of global poisonings, 56% of severe acute liver injury and acute liver failure, and 7% of drug-induced liver injury. We estimate that 0.4% of paracetamol poisoning cases are fatal.

• Access to paracetamol differs widely internationally, some countries allow unlimited non-pharmacy sales while some only allow sales in pharmacies. Available pack sizes also vary substantially.

• Some countries have introduced effective interventions aimed at reducing self-harm poisonings with paracetamol. This includes pack size restrictions in the UK and Denmark, and Denmark placing an age limit on sales.

Declaration of interest

A Chidiac is the recipient of a PhD stipend from Reckitt Benckiser as part of an untied educational grant awarded to R. Cairns. R. Cairns has also recieved conference speaker fees/honoraria from Reckitt Benckiser and The Pharmacy Guild of Australia. These funders had no role in the preparation of this review. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer Disclosures

One reviewer receives research funding on the safety of acetaminophen from the US sponsor (J&J), there are no personal financial relationships. The remaining reviewers have no other relevant financial relationships or otherwise to disclose.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/17425255.2023.2223959

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This paper was not funded.