Utilizing Mechanistic Biomarkers in Treating Paracetamol Hepatotoxicity

MANAGING PARACETAMOL OVERDOSE

PCM remains the leading global treatment for pain and fever and is typically sold over-the-counter in package sizes sufficient to cause ALF when ingested at once. Due to the narrow therapeutic bandwidth and unrestrained availability, PCM overdose accounts for approximately 100,000 annual hospital visits in the United Kingdom and roughly 50% of all ALF cases in the United States and Europe.2 The glutathione precursor N-acetylcysteine (NAC) has been the mainstay of treatment for the last decades, with no salvage therapy other than liver transplantation available in NAC-refractory cases. Despite the development of novel prognostic tools,3 the timely identification of high-risk cases remains difficult. Due to the rapid deterioration of liver function, a considerable number of patients with PCM-related ALF die while waiting for a new liver. These data have forced some to advocate the discontinuation of PCM altogether.2

Current treatment algorithms are based on the biochemistry of PCM toxicity. When present in excess, PCM is metabolized by the microsomal enzyme CYP2E1. PCM elimination by CYP2E1 produces the intermediate N-acetyl-p-benzoquinone imine (NAPQI), which binds and inactivates the antioxidant glutathione. At toxic PCM concentrations, NAPQI depletes glutathione reserves, causing oxidative stress and extensive hepatocellular necrosis in a matter of hours. In this therapeutic window, replenishing glutathione stores with NAC effectively prevents PCM hepatotoxicity. Based on the previous, the Rumack-Matthew treatment nomogram juxtaposes plasma PCM concentrations to the interval between PCM ingestion and hospital presentation to decide whether or not to initiate NAC treatment. The nomogram however does not predict the severity of liver injury. In addition, it cannot be used in unintentional or staggered overdose cases or in patients with an unclear medical history including abuse of alcohol. The current adage therefore is to treat with NAC when in doubt while monitoring liver biochemistry and to continue NAC treatment until transaminase levels have normalized. Given these uncertainties, more reliable tools to guide NAC treatment and identify high-risk cases are warranted.

BIOMARKERS AND RISK STRATIFICATION

To aid in risk stratification, Dear et al.1 collected data from two large cohorts of patients with a staggered, single dose, or unknown type of PCM overdose who required NAC treatment. They assessed the performance of five mechanistic biomarkers of PCM toxicity to predict the onset of liver injury (peak alanine aminotransferase [ALT] >100 U/L, primary outcome) or the onset of coagulopathy (international normalized ratio (INR) >1.5, secondary outcome). The cohort mainly comprised relatively benign cases, as the primary outcome was met in only ±11% of participants. In addition, a peak ALT >1000 U/L or an INR >1.5 was observed in just 2-3% of patients. None of these patients died or required liver transplantation. Patient comorbidity or drug use other than alcohol coingestion that could affect primary or secondary endpoints were not reported.

In patients with normal transaminases at presentation, the biomarkers miR-122, HMGB1, full-length K18, and caspase-cleaved K18 all (separately) predicted a peak ALT >100 U/L more accurately than plasma ALT or PCM concentrations at presentation. In both cohorts, the area under the curve of the receiver operator characteristic (ROC-AUC) was ≥0.93 for miR-122, full length K18, or HMGB1 versus ∼0.60 for systemic ALT or PCM levels at presentation. Similar results were found when considering a peak ALT >1000 U/L. Maximal performance was obtained with a model combining miR122, HMGB1, and the two K18 isoforms. In patients with normal ALT and INR at presentation, this set of biomarkers identified 49 of 50 patients who developed an ALT >100 U/L and recognized 824 of 825 of patients who did not develop liver injury. HMGB1 at admission best predicted the onset of coagulopathy. The ROC-AUC of 0.94 in the derivation cohort translated to positive and negative predictive values of 92.0 and 88.0, respectively. Slightly less favorable numbers were recorded in the validation cohort. For comparison, the ROC-AUC of ALT and PCM approximated 0.50 in both cohorts, indicating poor accuracy.

It is debatable how these biomarkers may improve the management of patients with PCM intoxication in the (near) future. Using biomarkers to decide whether patients require NAC treatment is likely restricted by assay turnaround times and current assay availability. The prognostic precision of these assays should also approximate 100% to safely withhold NAC treatment in fringe cases. Being able to predict a peak ALT >100 U/L may help to estimate the required duration of NAC treatment and, thereby, the length of hospital stay. Although helpful, this advance seems somewhat trivial when taking the ultimate consequence of PCM toxicity (i.e., ALF) into account. As there was no mortality or need for liver transplantation in the cohort, it remains to be demonstrated whether the tested biomarkers can also predict ALF. The low number of ALF cases may also reflect a drawback of robust research, as excluding patients unable to give informed consent may have biased the cohort towards the less severely ill.

FUTURE DIRECTIONS

Two phases can be distinguished in the progression from PCM overdose to ALF (Fig. 1). As mentioned, hepatocyte oxidative stress drives the earliest stages of PCM toxicity. NAC selectively acts on this stage by replenishing glutathione stores. Limiting the conversion of PCM to NAPQI could potentiate NAC treatment by diminishing the oxidative hit. The enzymes that convert PCM to NAPQI are controlled by hepatic nuclear receptors such as the constitutive androstane receptor (CAR) and the pregnane X receptor (PXR). In vitro studies have demonstrated that blocking CAR/PXR activity renders hepatocytes more resilient to otherwise toxic PCM levels.4 Risks of this approach are the concurrent shutdown glutathione synthesis from NAC. In NAC-refractory cases, oxidative stress translocates to mitochondria and causes cell death through the formation of protein inducts. The use of mitochondria-targeted antioxidants, which have become widely available, could help to limit hepatoxicity at this stage.

As many patients with PCM toxicity present late, a more appealing approach may be to target the sterile immune response downstream to hepatocyte necrosis. The chief consequence of necrosis is the extracellular release of damage-associated molecular patterns (DAMPs), which differ from ‘static’ injury markers (e.g., transaminases) in that they aggravate liver injury by mobilizing an immune response. Considering that immune cell-mediated liver injury triggers additional DAMP release (Fig. 1), only treating the initial hit with NAC at this point is insufficient to stop the progression of liver injury, in particular, because DAMP release may even precede the rise in plasma transaminases in some cases.1 In addition to HMGB1, systemic concentrations of DAMPs such as double-stranded DNA and mitochondrial DNA are also raised after PCM overdose.5, 6 The inflammatory pathways activated by these DAMPs have been elaborately studied. Accordingly, blocking leukocyte DAMP receptors such as toll-like receptor 9 has shown promise in preclinical models of PCM toxicity.7 The merits of neutralizing HMGB1 directly have also been demonstrated in animal studies.8 Expanding the therapeutic repertoire for PCM overdose beyond NAC therefore seems an attainable objective for future research. To that end, the clinical validation of animal studies is highly needed.