Acetaminophen (APAP), also known as paracetamol, is one of the most widely prescribed drugs for the management of pain and hyperthermia, even though its mode of action is still under investigation . The current maximum recommended dosage of APAP is 4 g/day in adults and 60–75 mg/kg/day in children. Although APAP is usually considered as a safe drug, APAP intoxication after overdosage can lead to massive hepatocellular necrosis and acute liver failure [2, 3]. At the moment, N-acetylcysteine is the only approved drug to treat APAP-induced liver injury in patients after voluntary, or unintentional, APAP overdose . Importantly, there is accumulating evidence that the maximum recommended dose can induce mild-to-moderate hepatic cytolysis, even in healthy individuals [5, 6]. It is also noteworthy that long-term APAP ingestion can induce chronic liver injury in some patients, such as granulomatous hepatitis and cirrhosis [7, 8].
Different predisposing factors are known, or suspected, to significantly enhance the risk and severity of APAP-induced hepatotoxicity. These factors include malnutrition and fasting, chronic alcohol abuse and alcoholic liver disease, and hepatitis C virus infection [9-11]. Another factor could be comedication with drugs inducing cytochromes P450 (CYPs), such as isoniazid, rifampicin and phenobarbital [12, 13], or with drugs which may modulate liver injury and/or repair . The risk of APAP-induced liver injury could also be enhanced by some gene polymorphisms, in particular in genes encoding APAP metabolizing enzymes [15, 16]. Lastly, a large retrospective study also reported that obesity-related nonalcoholic fatty liver disease (NAFLD) could be a risk factor for APAP-induced acute liver injury after APAP overdose .
Nonalcoholic fatty liver disease refers to the large spectrum of liver lesions in obese individuals ranging from fatty liver to nonalcoholic steatohepatitis (NASH) and cirrhosis. In some countries such as the United States, it is estimated that the prevalence of NAFLD in the general population could be about 20% in adults and 10% in children and adolescents [17, 18]. Actually, a majority of overweight and obese people develop fatty liver, which is characterized by the presence of large vacuoles of lipids (mainly triacylglycerol) within the hepatocytes. During obesity, a major mechanism leading to hepatic accretion of lipids is insulin resistance, which favours the redistribution of fat from adipose tissue to the liver and higher de novo lipogenesis in hepatocytes [19, 20]. Importantly, fatty liver can progress in the long term to NASH in 10–20% of patients. NASH is characterized not only by steatosis but also by necroinflammation (with ballooning degeneration of hepatocytes), some fibrosis and the presence of apoptotic hepatocytes. Key factors involved in the pathogenesis of NASH include oxidative stress, lipid peroxidation, mitochondrial dysfunction and overproduction of proinflammatory and profibrotic cytokines [20, 21]. Impaired activity of the mitochondrial respiratory chain and induction of CYP2E1 are two major factors leading to reactive oxygen species (ROS) overproduction during NASH [20, 22].
As previously mentioned, clinical investigations reported that NAFLD could favour APAP-induced acute liver failure after an overdose. Moreover, studies were performed in different animal models of obesity and NAFLD to find out whether these metabolic diseases could favour APAP-induced liver injury. To find out the clinical and experimental studies dealing with the topic of this review, we performed a PubMed search of literature published in the English language using the combination of at least two of the following queries: acetaminophen, obesity (or obese), liver, ‘fatty liver’, NASH, injury and ‘risk factors’. In the present article, the main findings of these clinical and experimental investigations will be presented and subsequently discussed. Prior to these sections, some major features of APAP metabolism and liver toxicity will be recalled.
Main features of APAP metabolism and hepatotoxicity
This section will only summarize the current knowledge about APAP metabolism and mechanisms of hepatotoxicity as several comprehensive reviews were recently published on these topics [23-26]. Briefly, APAP is mainly metabolized in the liver by phase II conjugating enzymes into the nontoxic glucuronide and sulphate conjugates, which respectively represent about 55% and 30–40% of the initial APAP dose [24, 27]. In addition, a small amount of APAP is oxidized to the highly reactive metabolite N-acetyl-p-benzoquinone imine (NAPQI) by several CYPs, especially by CYP2E1 and CYP3A4 [28, 29]. There is strong evidence that CYP2E1 is the main enzyme generating NAPQI in mice [29, 30], but several investigations suggested that CYP3A4 could also be an important enzyme catalysing NAPQI generation in human [31-33]. Importantly, obesity and related metabolic disorders are associated with significant alterations of APAP pharmacokinetics. In particular, increased hepatic activity of CYP2E1 and higher APAP glucuronidation have consistently been reported in obesity and NAFLD [22, 34-36], although the exact mechanisms of these alterations are still poorly understood. In contrast, hepatic activity of CYP3A4 is reduced in obesity and NAFLD [36, 37].
After its formation, NAPQI is rapidly detoxified by hepatic reduced glutathione (GSH) when APAP is taken at the recommended dosage. However, after APAP overdose, high levels of NAPQI can induce several detrimental effects within the liver. Indeed, hepatocellular GSH stocks are quickly depleted and once GSH is no longer available for NAPQI detoxication, this reactive metabolite binds to different macromolecules, including cytosolic and mitochondrial proteins [38, 39]. These early cellular events are then followed by c-jun N-terminal kinase (JNK) activation, profound mitochondrial dysfunction and ATP depletion, overproduction of ROS and peroxynitrite and massive hepatocellular necrosis [23, 26, 40]. Finally, this initial liver injury is followed by other key events such as inflammation, repair and regeneration that play a major role in the outcome of APAP-induced hepatotoxicity [23, 25, 41].
Data from clinical investigations
As already mentioned in the introduction, a large retrospective study reported an increased risk of APAP-induced acute liver injury after APAP overdose in patients with NAFLD . In this study, patients with pre-existent NAFLD and hospitalized for APAP overdose had more than a seven-fold higher prevalence of acute liver injury as compared to those without NAFLD . In a subsequent report, Myers and Shaheen reanalysed the same database by taking into account potential differences in overdose circumstance, which could be a significant confounding factor . This reanalysis supported the previous findings of a higher risk of APAP-induced hepatotoxicity in patients with NAFLD , even though the magnitude of this risk was attenuated (i.e. 3.91 vs. 7.43). Importantly, both studies showed that the risk of severe acute liver injury after APAP overdose was increased about similarly with NAFLD and alcoholic liver disease [9, 42]. In contrast, two studies reported that the susceptibility to APAP-induced acute liver injury was similar, or lower, in obese patients compared to nonobese individuals, but NAFLD was unfortunately not taken into consideration in these investigations [43, 44]. Interestingly, in one of these studies, obese patients had significantly poorer outcomes after acute liver failure . Altogether, these clinical investigations suggest that NAFLD, rather than obesity by itself, could increase the risk of APAP-induced acute liver injury after APAP overdose. This important point will be discussed later on. However, it is noteworthy that the discordant results between the aforementioned studies on the association between NAFLD/obesity and APAP-induced acute liver injury could be caused by differences in the criteria adopted to define these conditions in patients.
As previously mentioned, there is evidence that prescription of APAP at the recommended dosage can induce mild-to-moderate hepatic cytolysis, even in healthy individuals [5, 6]. However, it is still unclear whether NAFLD could enhance the risk, or the severity, of liver injury after therapeutic dose of APAP, although this has been reported in some obese patients .
Data from experimental studies
Different rodent models of obesity and NAFLD can be used experimentally, in particular in investigations dealing with the hepatotoxic effects of chemicals such as APAP (Table 1). Classically, obesity and related metabolic disorders can be induced by different types of high-calorie diets (including high-fat and high-sucrose/fructose diets), by genetic defects altering different key physiological functions such as food intake and energy expenditure, or by the combination of nutritional and genetic approaches [46-50]. Among the more commonly used murine models of genetic obesity, the C57BL/6J-ob/ob (ob/ob) are characterized by a mutation in the leptin gene. Indeed, lack of the adipokine leptin leads to increased food consumption and lower energy expenditure, thus resulting in massive obesity, insulin resistance, dyslipidaemia and severe fatty liver [51, 52]. C57BL/KsJ-db/db mice and Zucker fa/fa rats present a mutation in the gene encoding leptin receptor, which disrupts leptin signalling and induces obesity, insulin resistance and fatty liver [46, 52]. The KK-Ay mice, which present a defect in the signalling of the melanocortin receptors, spontaneously develop obesity, fatty liver, type 2 diabetes, arterial hypertension and diabetic nephropathy [46, 52, 53]. These genetic rodent models of obesity are associated with metabolic disorders that are more severe compared to the nutritional models, but usually with less variability between the animals.
|References||Animal models of NAFLD||Presence of NAFLD||Dose of APAP||Response to APAP toxicity||CYP2E1 activity|
|Corcoran and Wong ||Male Sprague-Dawley rats fed with a high-fat diet for 24 weeks||Not reported in this studya||710 mg/kg (i.p.)||Higher toxicity after 48 h, compared to rats fed a standard diet||Not reported in this studya|
|Blouin et al. ||Obese male Zucker fa/fa rats||Not reported in this studyb||1300 mg (p.o.)||Similar toxicity after 48 h, compared to lean rats||Not reported in this studyb|
|Ito et al. || |
Male C57Bl/6 mice fed a ‘Western-style’ diet for 16 weeks
Male ob/ob mice
Not reported in this studyc
300 mg/kg (p.o.)
300 mg/kg (p.o.)
Lower toxicity after 6 h, compared to mice fed a standard diet
Lower toxicity after 6 h, compared to wild-type mice
Not reported in this studya
Not reported in this studyc
|Kon et al. ||Male KK-Ay mice||Yes||300 or 600 mg/kg (p.o.)||Higher toxicity after 6 h, compared to wild-type mice||Not reported in this studyd|
|Kucera et al. ||Male Sprague-Dawley rats fed a high-fat diet for 6 weeks||Yes||1 g/kg (p.o)||Higher toxicity after 24 and 48 h, compared to rats fed a standard diet||Not reported in this studya|
|Aubert et al. || |
Female db/db mice
Female ob/ob mice
500 mg/kg (p.o.)
500 mg/kg (p.o.)
Higher toxicity after 8 h, compared to wild-type mice
Similar toxicity after 8 h, compared to wild-type mice
|Donthamsetty et al. ||Male Swiss Webster mice fed a MCD diet for 1 month||Yes||360 mg/kg (i.p.)||Higher toxicity from 6 to 48 h after overdose, compared to mice fed a standard diet||Unchanged|
From 1987 to 2012, six different studies reported the effect of acute APAP intoxication in several animal models of obesity and NAFLD [54-59]. The main data from these investigations are presented in Table 1. In addition, one study reported APAP hepatotoxicity in a murine model of methionine and choline deficient (MCD) diet-induced steatohepatitis (Table 1) . However, it must be pointed out that the MCD diet consistently leads to a dramatic loss of body weight [61-63] and does not induce systemic insulin resistance [63, 64]. It is also noteworthy that the doses of APAP greatly varied in these seven studies, with the highest doses used in the rat models (Table 1). This is because rats are much more resistant to APAP-induced liver injury compared to mice [65, 66].
Among these different studies (Table 1), four of them showed that NAFLD was associated with higher liver injury after a single APAP overdose [57-60]. Another study also showed higher APAP toxicity in rats fed a high-fat diet , although the presence of NAFLD can only be suspected in this study (Table 1). In contrast, the remaining investigations showed similar  or lower liver injury . Unfortunately, there is no definite explanation for these discrepancies, although some hypotheses can be put forward. For instance, investigations in male and female ob/ob and db/db mice showed that the hepatic CYP2E1 status could play a significant role . Indeed, higher APAP hepatotoxicity was observed in female db/db mice but not in female ob/ob mice and this was associated with a specific induction of CYP2E1 activity in the first group of animals. Interestingly, hepatic CYP2E1 activity was unchanged  or reduced  in male ob/ob mice and this could explain why Ito and collaborators reported lower acute APAP hepatotoxicity in this model . It is also noteworthy that plasma levels of APAP-glucuronide and APAP-sulphate were almost similar between female db/db mice and female ob/ob mice after 500 mg/kg of APAP . Hence, the observed difference of APAP-induced hepatotoxicity between these murine models cannot be attributed to APAP glucuronidation and sulphation.
At the moment, it is still unclear why hepatic CYP2E1 is specifically induced in female but not in male db/db mice . Although there was an overall positive correlation between hepatic CYP2E1-mediated aniline hydroxylase activity and plasma glucose in all groups of untreated mice (i.e. wild-type, ob/ob and db/db of both sexes), plasma glucose levels were not different between male and female db/db mice . Thus, hepatic CYP2E1 induction in female db/db mice may not be explained by the severity of insulin resistance and type 2 diabetes in these animals, contrary to what has been suggested by some investigations . In addition, there was no difference in plasma β-hydroxybutyrate levels between female and male db/db mice . This suggests that ketone bodies may not be responsible for CYP2E1 induction in female db/db mice, although previous investigations found a significant correlation between plasma β-hydroxybutyrate levels and hepatic CYP2E1 activity [34, 68]. Thus, further investigations will be needed to determine the exact mechanism of higher hepatic CYP2E1 activity in female db/db mice.
Besides differences in APAP biotransformation to NAPQI, higher susceptibility to APAP-induced liver injury might be explained by lipid accumulation and oxidative stress associated with NAFLD. Regarding lipid accretion, it is noteworthy that higher APAP hepatotoxicity in female db/db mice compared to female ob/ob mice was observed despite lower lipid accumulation in db/db liver . Importantly, the basal plasma levels of ALT and AST (i.e. before APAP intoxication) were comparable between female db/db and female ob/ob mice . Thus, increased susceptibility to APAP hepatotoxicity in female db/db mice cannot be attributed to the pre-existence of higher hepatic cytolysis. However, further studies will be required to find out whether the accumulation of some lipid species could be involved. Indeed, it is noteworthy that db/db hepatocytes present more microvesicular steatosis and less macrovacuolar steatosis compared to ob/ob hepatocytes [50, 59]. Because the size of the droplets is greatly dependent on their lipid composition [69, 70], it is possible that the accumulation of some specific lipid species in female db/db hepatocytes could have played a role in the higher susceptibility to APAP toxicity.
Nonalcoholic fatty liver disease is associated with oxidative stress, which can alter different cellular components such as nucleic acids and unsaturated fatty acids [20, 71, 72]. Moreover, hepatic GSH reserves could be lowered in NAFLD, thus reducing the antioxidant capacity of the fatty hepatocytes [20, 73]. Hence, the higher susceptibility to APAP hepatotoxicity observed in some rodent models of obesity might be related to basal oxidative stress and lower GSH levels. Yet, basal hepatic GSH levels were not significantly reduced in two rodent models of obesity showing higher vulnerability to APAP toxicity [58, 59]. It is however noteworthy that total GSH levels were assessed in these studies and thus it will be important to determine whether mitochondrial GSH was specifically depleted. Indeed, some investigations suggested a key role of the mitochondrial pool of GSH in APAP-induced liver injury [74, 75].
As previously mentioned, JNK activation plays a major role in APAP-induced hepatotoxicity [23, 26, 40]. Interestingly, JNK activation could also play a significant role in NAFLD pathogenesis [76, 77]. Only two studies examined the hepatic JNK status in obese animals intoxicated by APAP [57, 59]. In the first study, JNK activation (as assessed by phospho-JNK) after APAP intoxication was significantly stronger in obese and diabetic KK-Ay mice compared to wild-type mice, and this was associated to higher APAP hepatotoxicity in the obese animals . In the second study, increased APAP hepatotoxicity in db/db mice was not associated to higher JNK phosphorylation . Unexpectedly, hepatic JNK was specifically activated in ob/ob mice although these animals did not present higher APAP liver injury . Thus, JNK activation might not be a key mechanism of APAP hepatotoxicity in some murine models of NAFLD. Further investigations will be needed to assess phospho-JNK not only in whole liver but also in isolated mitochondria .
In addition to the investigations performed in rodent models of obesity, one study reported the effect of acute APAP intoxication in a murine model of MCD diet-induced steatohepatitis (Table 1) . In this study, higher APAP hepatotoxicity was observed despite the lack of pre-existent CYP2E1 induction and the presence of normal basal hepatic GSH levels , suggesting the involvement of other mechanism(s). Interestingly, the basal hepatic ATP levels were already reduced by 45% in mice fed the MCD diet , which is consistent with a significant impairment of the mitochondrial respiratory chain and oxidative phosphorylation in this experimental model [79, 80]. Hence, severe pre-existent mitochondrial dysfunction might have played a role in the higher APAP hepatotoxicity observed in mice fed the MCD diet.