Acetaminophen Toxicity: A History of Serendipity and Unintended Consequences (2025)

Acetaminophen (N-acetyl-p-aminophenol [APAP]), which is widely known in the United States by its trade name, Tylenol, that is derived from the underlined strings in the chemical name (N-acetyl-p-aminophenol) and as paracetamol in Europe (from its alternative chemical name para-acetylaminophenol), occupies a unique position in pharmacology and medicine: an over-the-counter drug, sold in billions of units annually in North America alone, but a powerful dose-related toxin.¹ Over the past 40years, APAP toxicity has accounted for 46% of all acute liver failure (ALF) in the United States² and between 40% and 70% of all cases in Great Britain and Europe.^3-5 APAP toxicity dwarfs by several fold the number of cases of ALF resulting from all prescription drugs combined (Fig. 1), and it has been the subject of two US Food and Drug Administration (FDA) Advisory Committee meetings in the past 15years. Even though spontaneous (i.e., transplant-free [TFS]) survival from APAP ALF is very high (approximately 70%) compared with all other prescription drugs, alternative medications, dietary supplements, and recreational compounds (approximately 30% TFS), the very high incidence of APAP toxicity causing ALF means that deaths from this etiology are almost double those from other xenobiotics. The history of APAP is unique and provides insights into how happenstance plays a role in medical breakthroughs.

History

APAP was initially discovered in 1889 in United Kingdom, recognized as a metabolite of phenacetin in 1949 and noted to lack any renal injury with chronic use. Paracetamol came into clinical use in Britain as Panadol in 1956.⁶ A decade later, estimates placed its sales in the United Kingdom alone at 680million units per year.⁷ Thus, the popularity of paracetamol increased rapidly, possessing few of the side effects associated with phenacetin or aspirin (acetylsalicylic acid [ASA]), although the latter was still much more popular then. An ASA overdose was a thoroughly unpleasant affair, both for the patient, who suffered from intense tinnitus, impaired hearing, hyperventilation, vomiting, diplopia, dehydration, and faintness, and for the unfortunate, often junior physician who had the daunting task of salvage in the early hours of the morning in the emergency department. In 1966, a number of case reports associating paracetamol with severe liver injury and reports of fatal outcomes began to appear.^6-10 In 1966, Boyd and Bereczky⁸ discovered that paracetamol caused dose-related toxicity in rats. Each clinical case published during this early period described at least one recognized aspect of APAP toxicity: massive quantities (50 or more) of tablets consumed, rapid-onset illness, very high aminotransferases, mild hyperbilirubinemia, hypoglycemia, centrilobular necrosis, and full recovery for those who did survive. By 1970, paracetamol was the most frequently used suicidal agent in the UK, occasioning a report of 41 cases.¹¹ The first case report in a North American patient was that reported by Boyer and Rouff¹² in Journal of the American Medical Association in 1971. In 1972, the Liver Unit at King’s College Hospital London established the first two-bed Liver Failure Intensive Care Unit, which was frequently filled with young women on life support following attempts at self-harm.

Over the years leading up to 1973, when they published their seminal work,¹³ Jerry Mitchell and David Jollow were serving as Fellows in the famed Brodie Laboratory of Chemical Pharmacology (LCP) at the US National Institutes of Health (NIH) and had been involved in ongoing studies on the role of hepatic metabolism in the liver injury due to bromobenzene.

The resulting four articles by Jollow, Mitchell, Gillette, and others at the LCP in 1973^13-17 established the mechanism once and for all. Dose-related hepatic injury manifested as centrilobular necrosis: the higher the dose, the greater the area of injury (Fig. 2). This occurred when the capacity of sulfation and glucuronidation was exceeded, leading to metabolism by cytochrome P450 (CYP; primarily CYP2E1) and, with depletion of GSH, binding of the highly reactive metabolite to cell proteins via sulfhydryl linkage (Fig. 3). NAC turned out to be a suitable antidote.¹⁸ Oral NAC (Mucomyst) came into common usage within a few years and intravenous NAC shortly thereafter, although not until considerably later in the United States, in 2004,¹⁹ because the efficacy of oral NAC treatment was so well established and there were some adverse effects, admittedly few and modest, of intravenous NAC. Notwithstanding, there had been a dire need for an intravenous option for patients with severe APAP who could not tolerate oral NAC.

Paracetamol Crosses the Atlantic, Becoming Acetaminophen

APAP use was virtually nonexistent in the United States until the early 1980s, when, after the association of aspirin with Reye’s syndrome in children was reported,^{20, 21} APAP was seen as a suitable alternative and became marketed actively as Tylenol, as well as other brands. This was followed by development of convenience combinations, such as APAP/diphenhydramine (Tylenol PM, Nyquil, and others), as well as opioid/APAP combinations. APAP’s popularity increased dramatically, despite the fact that virtually all Reye’s cases were confined to children, not adults.²¹

Therapeutic Misadventures

In the late 1970s and early 1980s, numerous reports in the United States surfaced regarding severe liver injury associated, not with suicide attempts, but unintentional misuse.^22-24 These were inadvertent overdoses in the setting of acute or chronic pain, often accompanied by alcohol use and without suicidal intent. A single-time-point APAP overdose of 12 to 15g (24-30 “extra strength” [500-mg] tablets) was associated with an approximately 50% mortality rate if untreated; however, ingestions of 6 to 7g/day over several days in the right setting (alcohol/starvation) could yield the same result, a very similar hyperacute injury of great severity. Over the next decade, North American hepatologists became increasingly aware of this entity. A review of ALF in 1993 included mention that APAP was fast becoming the most frequent cause of ALF in the United States.²⁵ Zimmerman and Maddrey²⁶ published a comprehensive article in 1995 describing 67 cases of inadvertent toxicity, “therapeutic misadventures,” sometimes referred to as “the alcohol-Tylenol syndrome” because of the frequent association of alcohol use. A comprehensive review of APAP-related toxicity at a single large urban hospital substantiated the prior claims: 71 hospital admissions for APAP toxicity (not necessarily with liver failure) were identified over a 40-month period.²⁷ Criteria were established to distinguish the intentional (suicide) from the unintentional (therapeutic misadventure) phenotype (Table 1).²⁷

With the intentional group, many patients present to an emergency department early and receive NAC during the window of therapeutic response, resulting in more favorable outcomes, that is, before the “lethal” binding of NAPQI to hepatic protein targets is extensive. The unintentional group, consuming many products containing APAP to self-medicate conditions such as the “flu” or chronic pain, typically would be unaware that each preparation contained the drug, and hence that their combination provided maximal, or greater than maximal, therapeutic doses. Being unaware of the danger, presentation to an emergency department could be delayed until after the window of optimal NAC effect had passed and the symptoms of FHN were apparent. A history of chronic alcohol abuse and/or chronic usage over weeks seemed to exacerbate the problem, although many suicidal patients also are substance abusers.

In the frank overdose (intentional) situation, the mechanism originally observed in experimental animals, and in particular the crucial roles of the GSH threshold and of NAC as an antidote, has been directly applicable to understanding and treating human APAP overdoses. The mechanism underlying liver injury in the unintentional group is less clear. Although the total accumulated dose of APAP in these patients may be similar to the single dose of suicidal patients, the fact that the drug is consumed over days or weeks would suggest that the “daily” pressure on the GSH threshold would be much less than in the frank overdose situation. Why then is APAP hepatotoxic in these patients?

The results of animal studies do provide some suggestions. As was observed in the original difficulty in establishing an animal model at NIH, the glycogen status of the rats was crucial. Subsequent studies in animals confirmed that low glycogen levels result in a lower rate of production of uridine 5′-diphosphate-glucuronic acid, the essential cofactor for the glucuronidation of APAP. Glucuronidation is the major pathway of clearance of APAP in both laboratory animals and humans. Suppression of this major pathway results in a longer half-life and a greater fraction of the dose being converted to NAPQI. Suppression of the sulfotransferase pathway has a similar, although less profound, effect. It is well known that chronic abuse of ethanol causes enhanced activity of CYP2E1 in the liver, that is, the CYP most important in the metabolic activation of APAP to NAPQI.

Collectively, we can imagine a combination of factors that might enhance the susceptibility of, say, a patient unintentionally self-medicating excessively for “flu” and/or prolonged decreased interest in and consumption of food that could result in a significant decrease in glycogen and sulfate stores. Corresponding decreased clearance of APAP by the major nontoxic pathways (glucuronidation and sulfation) would provide enhanced chronic “daily” NAPQI pressure on the GSH-protective mechanism. A history of alcohol abuse causing enhanced hepatic CYP2E1 activity and NAPQI formation would further increase the pressure on the GSH threshold. One can imagine an accumulation of toxic “hits” over time until an injury threshold is exceeded and cell death occurs. However, the key piece of information to support this scenario is missing; namely, what is the GSH level in the liver under these conditions? The human body has ample reserves of cysteine (Cys) and methionine. Protein breakdown in the presence of dietary restriction should still provide ample support for GSH resynthesis. In rats, chronic extreme thiol-amino acid–deficient dietary studies indicate conservation of available Cys for the support of GSH synthesis. It is reasonable to expect a similar conservation in humans, but data are lacking.

APAP-CYS Adducts

Over the 30years following the 1973 NIH studies, additional details surrounding the pathogenesis of APAP-induced hepatic necrosis became clear.

The product of the covalent binding of the reactive intermediate, NAPQI, first recognized in 1973, was then determined to be bound to cysteine on cell proteins (Fig. 3).³² Using radiolabeled APAP, mouse liver homogenates that were exposed to the drug produced a product that could readily be identified by radioimmunoassay and later by high-pressure liquid chromatography with electrochemical detection (HPLC-ED). These were the APAP-Cys adducts (NAPQI bound via cysteine to cell proteins) that appear in tissue and plasma as early as 2hours after a high-dose exposure (Figs. 4 and 5).³³

Because radiolabeling is not feasible in patients, an antibody-based method specific for the adducts was used next to demonstrate adducts present in centrilobular areas of necrosis.^{34, 35} The first report of adducts in human serum was by Hinson et al.,^{36, 37} using patient sera collected by Henrik Poulsen in Denmark. However, HPLC-ED had greater sensitivity than the original antibody-based assays allowing “dose-response” studies across a range of exposures, from “low dose or therapeutic exposures” to overdoses associated with massive liver injury. These studies performed over many years by Jack Hinson, Dean Roberts, and Laura James at the University of Arkansas Medical Sciences, Little Rock, Arkansas, demonstrated that APAP-Cys adducts above a “toxicity threshold” corresponded to alanine aminotransferase (ALT) elevations >1000IU/L, proof positive of significant liver injury. APAP-Cys adducts are considered a reliable biomarker for APAP injury and remain in plasma for up to 10days following injury, as demonstrated initially in 2006³⁸ using HPLC-ED and later in larger studies with sera from adults and children in various overdose scenarios.^39-42

The next step was to develop a rapid assay for APAP-Cys, using monoclonal antibodies as a point-of-care assay.⁴³ This is still in development and will add immeasurably to our ability to diagnose APAP toxicity in the emergency department and intensive care units at all stages of its evolution. Serendipity, good timing, thoughtful investigation following leads, and many hands (Fig. 6) all contributed to the relatively rapid discovery of the APAP pathway and its antidote, NAC, which has saved so many lives over the years.

Yet, for all this progress in unraveling the cellular and chemical intricacies of the mechanisms by which the metabolism of APAP leads to hepatocyte necrosis, the story of APAP hepatotoxicity has not ended. It has become apparent in recent years that there are other cellular players involved, namely, the reactions of the innate immune system. The complexity of the downstream effects induced by APAP-related necrosis affect several intrinsic systems, including apoptosis, autophagy, and stress to the endoplasmic reticulum.^44-47 Because disruption of the innate response in some settings alleviates APAP hepatotoxicity in experimental animals independent of the production of NAPQI and its effects, perhaps there are opportunities to cash in therapeutically to prevent human APAP hepatotoxicity.

Current Status

By 1995, APAP had become the most common form of ALF in the United States,^{48, 49} having gained that distinction in Europe two decades earlier. In 1998, the British Houses of Parliament passed legislation limiting package size and using blister packs to deter impulsive behavior.^50-52 Whether this has proved effective has been a subject of continued debate.^{53, 54} Over the past three decades, little specific progress has been made in curbing these unfortunate events, other than prompt use of NAC. However, circumstantial evidence suggests that overall outcomes for those with severe liver injury may be somewhat better overall in the current era than a quarter century ago, in part due to better intensive care and in part due to availability of transplantation.⁵⁵

APAP-related hepatotoxicity still exceeds all other forms of ALF by several fold. Having available an antidote such as NAC has likely saved thousands of lives—indeed, APAP toxicity might have been a major scourge otherwise—or, absent an antidote, APAP might have been banned completely long ago. Two FDA advisory committees (2002 and 2009) sought to limit APAP-related toxicity,⁵⁶ mostly by attacking the problem around the edges: improved package labeling and efforts to limit the unintentional overdosing by limiting combination opioids to 325mg APAP per tablet. Thus far, little evidence of effect has been observed, but perhaps more time is needed before an effect will become discernible.⁵⁵ Public awareness remains modest, and research suggests that both the intentional and unintentional overdose patients suffer from impulsivity, that is, the tendency to make decisions that are not in one’s best interests.^{57, 58} These actions may actually have a genetic basis, as is supported by research in the genomics of addictive behavior.⁵⁹ Still, the ubiquitous presence of a dose-related toxin as an over-the-counter analgesic continues to baffle hepatologists who are all too aware of the problem. But understanding the risks is not new; an editorial in Lancet in 1975, 45years ago, opined: “Surely it is time to replace paracetamol with an effective analogue which cannot cause liver damage.”⁶⁰