Idiosyncratic DILI and RUCAM under One Hat: The Global View

: Drugs are prescribed worldwide to treat diseases but with the risk of idiosyncratic drug-induced liver injury (iDILI). The most important difﬁculty is how best to establish causality. Based on strong evidence and principles of artiﬁcial intelligence (AI) to solve complex processes through quantitative algorithms using scored elements, progress was achieved with the Roussel Uclaf Causality Assessment Method (RUCAM) in its original and updated versions, often viewed now as the gold standard. As a highly appreciated diagnostic algorithm, the RUCAM is in global use with around 100,000 iDILI cases published worldwide using RUCAM to assess causality, largely outperforming any other speciﬁc causality assessment tool in terms of case numbers. Consequently, the RUCAM helps to establish a list of top-ranking drugs worldwide implicated in iDILI and to describe clinical and mechanistic features of iDILI caused by various drugs. In addition, the RUCAM was recently applied in iDILI cases of patients treated for coronavirus disease 2019 (COVID-19) infections or cancer patients treated with immune checkpoint inhibitors (ICIs), as well as in the search for new treatment options with conventional drugs in iDILI. Analyses of RUCAM-based iDILI cases are helpful to support pathogenetic steps like immune reactions, genetic predisposition as evidenced by human leucocyte antigens (HLA) genotypes for selected drugs, and the role of the gut microbiome. To achieve consistency in data collection, analysis, and speciﬁc clinical and pathogenetic presentation, researchers, regulatory agencies, and pharmaceutical ﬁrms should place iDILI and the updated RUCAM as the causality tool under one and the same hat in review articles and clinical guidelines for the diagnosis and treatment of iDILI.

Diagnosis of iDILI is much more complex due to low case numbers that limit clinical experience, the large number of offending drugs, polymedication, heterogeneity of liver injury pattern, variability of therapeutic options, lack of animal models that could help identify underlying mechanistic steps, and different opinions regarding the causality assessment of iDILI [1,4,[13][14][15].The disease complexity requires expertise in the clinical field and enthusiasm to solve complex issues.Fascinating approaches are getting iDILI cases with their multifaceted clinical features under one hat with a globally applied diagnostic algorithms to firmly diagnose iDILI.
In this review article, we provide evidence that publication of iDILI cases systematically needs the use of RUCAM (Roussel Uclaf Causality Assessment Method) to firmly establish the diagnosis and concomitantly exclude alternative causes.This also helps describe clinical and mechanistic characteristics of iDILI and provide evidence-based pharmacotherapeutic options.

Search Terms and Strategy
The literature search strategy involved the PubMed database and Google Science, using the following terms: idiosyncratic drug-induced liver injury, RUCAM, updated RUCAM.For the term idiosyncratic drug-induced liver injury, 48,700 articles were provided if combined with the term RUCAM, and 63,000 articles if combined with the term updated RUCAM.Then, the first 50 publications derived from each term group were checked for their suitability to be included in this review article and provided the primary base for further analysis.The search was performed first on 30 April 2023 and then finalized on 25 May 2023.Publications were complemented by the large private archives of the authors.Limited to publications in the English language, there were no other restrictions regarding year of publication or study design.

RUCAM, Its Global Use and High Appreciation
In 1990, the definitions and some criteria for causality assessment of DILI were established and then finalized by an international consensus meeting with worldwide DILI experts as participants that included J.P. Benhamou (France), J. Bircher (Germany), G. Danan (France), W.C. Maddrey (USA), J. Neuberger (UK), F. Orlandi (Italy), N. Tygstrup (Denmark), and H.J. Zimmerman (USA) [16].These experts co-evaluated DILI cases for case characteristics, diagnostic criteria, liver injury pattern, and reexposure criteria.They standardized DILI case assessment with specific and quantitative items, and they all received appropriate credits for their ambitious contribution [16].Later, in 1993, partially based on the results of the consensus meeting that were not yet a scoring system, the original RUCAM was developed and validated by the team from Roussel Uclaf [17,18].RUCAM was a new causality assessment method (CAM) intended to overcome experts' and clinicians' previous problems with unstructured evaluations lacking defined and scored items resulting in debated causality key elements, definitions of terms related to liver injuries, and chronological criteria [17,18].In 2016, the original RUCAM was updated [19].
Since 1993, and until mid-2020, 81,856 DILI cases with increased annual numbers were published following causality assessment using the original or the updated RUCAM [20], outperforming in terms of case number any other published method listed earlier [19,[21][22][23][24].Among the top-ranking countries with the most DILI cases assessed by the international consensus criteria of RUCAM were China, the United States, Germany, Korea, and Italy (Table 1) [20].
China being in place #1 (Table 1) [20] could be due to the large population and likely the 2017 Chinese guidelines for the diagnosis and treatment of drug-induced liver injury, published by the Drug-induced Liver Injury (DILI) Study Group, the Chinese Society of Hepatology (CSH), and Chinese Medical Association (CMA) [24], enforcing the use of the updated RUCAM for DILI cases [19].The US being in rank #2 (Table 1) [20] is based on 32 publications compiled in a separate listing (Table 2) .The use of RUCAM by authors from the US (Tables 1 and 2) is interesting and highly appreciated as several of them are known members of the US DILI network (Table 2) [20].They used RUCAM by switching from their own network method [58] to the universally used RUCAM [17,19], although RUCAM was not promoted by the network [58].This strong signal by US DILI experts for the use and appreciation of RUCAM is understandable in view of the long list of methodology shortcomings of the network method described in detail and published [58], which are partially provided in a list (Table 3).
It seems that the difficulties inherited in the network method were truthfully reported (Table 3) [58], making the method inadequate for daily and global use.The lack of proper validation of this method using positive results of reexposure tests is common with many other methods published in recent years as summarized [22,23].Data derived from CAMs not properly validated must be viewed with caution and should not be published.This shortcoming also applies to CAMs based on subjective opinion and not on standard liverinjury-specific elements with individual scores and definitions (Table 3) [58].Under these conditions, qualifying the network method as standard [58] is certainly disputable, also in comparison with RUCAM and its known qualities [1,[17][18][19][20].  [20].

Experiences and Weaknesses of the US DILI Network Method [58]
• Cases were enrolled in the registry within 6 months of DILI onset and underwent global introspection with so-called expert opinion • Causality assessment in real time for clinicians' use was not feasible

Experiences and Weaknesses of the US DILI Network Method [58]
• There was no accepted definition provided for an expert in DILI • For each case, consensus must be achieved excluding minority votes • Consensus is still a subjective opinion • The network process restricts the naming of offending agents to 3 • Strong opinions or biases of a single experts were reported • Lengthy and lively conversations often occurred during the processes • The network process is described as cumbersome, time-consuming, and costly, needing data exchanges, monthly meetings, and logistics with administrative, organizational, and technological expertise • Each case received a final likelihood range as percentage, arbitrarily given by the assessors not based on individually scored elements • Total bilirubin was one of the inclusion criteria if >2.5 mg/dl without ruling out unconjugated hyperbilirubinemia due to, e.g., Gilbert syndrome • Network experts missed the diagnosis of HEV in wrongly diagnosed DILI cases needing a downgrading of percentage DILI likelihood • Not using a gold standard, a good method reliability was assumed • External validation of the method with a different group of experts is explicitly discouraged as labor is considered intensive and expensive • The network method was used only in US centers • Despite the weaknesses, the network method is assumed as best standard for the time being, but it was imperfect in 2016, asking for mandatory improvements • Finally, the original RUCAM was surprisingly quoted and described with 11 plain words: "RUCAM requires decline in liver enzymes to get a high score".
• Assessment is user friendly, cost effective with results available in time, and without need of expert rounds to provide arbitrary opinions [1,19,21,23,63] • Transparency of case data and clear result presentation [1,19,21]

RUCAM with Its Basic Features and Specifics
• Suitable for reevaluation by peers [1] and regional registries, national or international regulatory agencies, and pharma firms [1,19,64] • Mandatory use to validate future diagnostic biomarkers [65,66] and HLA association • Encourages prospective case data collection to obtain best results; however, RUCAM is also prepared for studies with a retrospective study protocol [19] • Real-time evaluation of the DILI case at the bed side [19] Clearly defined and scored key elements [19] • Time frame of latency period

• Unintentional reexposure
Other important specifics [19] • Laboratory-based liver injury criteria Advantages of RUCAM include the incorporation of positive results obtained after unintentional reexposure in its diagnostic algorithm (Table 4).If these data are available, a score of +3 is added, which helps achieve total scores of highly probable causality gradings [19]; however, high gradings are achievable even in the absence of reexposure results.An agreement exists that intentional reexposures are discouraged due to high risks of acute liver failure.The inclusion of reexposure data requires the application of criteria according to clear definitions (Table 5) [19].

Challenges and Limitations of RUCAM
• The quality of published RUCAM-based case data depends strongly on the qualification and experience of the submitting physician • RUCAM cannot compensate for inadequate quality data and case providers not familiar with liver diseases; quality problems also remain on the side of the reviewers and journal management [67][68][69][70] • Intentional upgrading of causality levels from possible to probable in cases initially assessed by the objective updated RUCAM and subsequently reassessed by the global introspection in a report with western co-authors remain debatable [66] as substantiated in three Letters to the Editor, presented by authors from India and Iceland [68], and China [69,70] • Fraudulent upgrading from possible to probable RUCAM gradings of published cases with the intention to provide more power on risky liver injury, uncovered in court, is outside of any ethics standard [71] • Challenging are reports entitled as DILI, but in fact several cohorts were lumped together with non-drugs like herbs or so-called dietary supplements as causatives of HILI, providing biased results for drugs and the other causatives due to cohort heterogeneity • Publications occasionally report on RUCAM-based DILI cohorts that include cases with a possible causality grading, which confounds good data with a probable or highly probable causality level [66].This problem must be solved prior to submission by deleting all cases with a possible or lower causality grading from the analysis to be published • Challenging for RUCAM are mixed cohorts of DILI caused by multiple medicinal products without providing individual RUCAM scores for each product, or giving causality gradings as means ± SEM or ± SD for drug groups [20] • Misuse of RUCAM by using reports on DILI without values of ALT and ALP, preventing both, verification of criteria characterizing the liver injury as well as calculation of the R (ratio) and selection of the appropriate RUCAM subtype for correct causality assessment [19] • Misuse of RUCAM by attempting including results of positive unintentional reexposure without adherence to the specific criteria [19] Abbreviations: ALP, alkaline phosphatase: ALT, alanine aminotransferase; DILI, drug-induced liver injury; HILI, herb induced liver injury; RUCAM, Roussel Uclaf Causality Assessment Method.

Validation
Among many positive aspects, special attention was paid to the good results of the validation method of RUCAM, which can be independently reassessed by different colleagues due to the clear criteria identification, stepwise approach, and easy scoring system [1].This appreciation substantiates previous reports on this topic, published in 1993 [17,18], 2005 [59], 2019 [61], and more recently in 2022 [60], as evidenced by additional details (Table 7) [59][60][61].

Reports on Validation of RUCAM
• RUCAM was internally validated using published DILI reports with positive test results of reexposure, also named positive rechallenge, which demonstrated without incorporation of the rechallenge test into the score, a sensitivity of 86%, specificity of 89%, positive predictive value of 93%, and negative predictive value of 78% [18], providing results that were commonly appreciated [1] and underlined the value of the original RUCAM as robust diagnostic algorithm [17].Positive unintentional reexposure tests are considered as gold standard among DILI experts [1,18], as erroneous reexposure of a suspected drug provides in retrospect the strongest evidence for DILI [1] if strict criteria were fulfilled [19].The good validation data were confirmed by subsequent studies [59][60][61].
• A good reliability based on interrater agreement by using the original RUCAM for DILI cases was reported as first external study [59].
• A second external study reported that there were no discrepancies in assessment by the two hepatologists who used the original RUCAM in suspected iDILI cases due to sevoflurane and desflurane [61].This was a prospective incidence study of fifteen patients that provided RUCAM-based causality gradings of highly probable in 3 cases, probable gradings in 5 cases, and possible gradings in 7 patients.
• A third external validation study used the updated RUCAM for determination of causality described in 72 patients with COVID-19 and suspected DILI [60].Two independent rating pairs (consisting of two clinical pharmacologists plus 2 general physicians), who had received a short training program for pilot testing just prior to the actual RUCAM use, determined the likelihood of DILI using the RUCAM scale in these DILI patients.As a result, the overall Krippendorf kappa was 0.52, with an intraclass correlation coefficient (ICC) of 0.79 and viewed as excellent reliability for using the updated RUCAM [60].Whether this is achieved through the prior training remains to be verified by assessors without prior training.Confirming previous reports [18,59], this good reliability result was remarkable as it was based on a retrospective study design [60].

Liver Injury Criteria
RUCAM criteria of iDILI are defined as serum activities of ALT (alanine aminotransferase) or ALP (alkaline phosphatase) with ALT ≥5 times the upper limit of normal (ULN) and/or ALP ≥2 times ULN.Both enzymes are diagnostic parameters of liver injury, whereas total bilirubin, as a parameter of liver function when conjugated bilirubin is predominant, is explicitly not part of the diagnostic RUCAM algorithm [19].In addition, unconjugated hyperbilirubinemia, which has nothing to do with DILI, can be seen in patients with the genetic Gilbert syndrome and a frequency of up to 8% in the general population [72].Cases with ALT or ALP below the thresholds are not considered as liver injuries that could carry a risk for patients but are categorized as liver adaptation or liver tolerance [73].

Liver Injury Pattern
RUCAM is also appreciated for taking into consideration the liver injury pattern, defined in 1990 [16] as hepatocellular injury, cholestatic liver injury, or mixed liver injury, first published in 1993 [17] and mentioned again in 2016 [19], and now shown in a flow chart (Figure 1).The determination of the individual liver injury pattern is a prerequisite for the causality assessment of suspected DILI cases by the updated RUCAM that exists with two versions, one is destined for the hepatocellular injury, and the other one for the cholestatic or mixed liver injury.The approach is identical for suspected herb-induced liver injury (HILI).Adapted from a previous open-access publication [19].Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; DILI, drug-induced liver injury; R, ratio; and ULN, upper limit of normal.
This classical differentiation of liver injury pattern, known also as phenotypes, is used in most DILI reports, although occasionally without quoting the source.It is mandatory for a causality assessment to use RUCAM but also helpful for the description of clinical DILI features.As the determination of the liver injury pattern requires only the results of serum ALT and ALP activities (Figure 1), this approach saves financial resources and does not require an invasive and risky liver biopsy.

Top Drugs
Several publications presented listings of drugs most implicated in iDILI, with differences due to preferred drug treatment that may vary among populations.Examples are anti-tuberculosis drugs that rank high in countries with many patients treated for tuberculosis.

Global Analysis
Top-ranking drugs causing iDILI assessed by RUCAM were retrieved from various publications throughout the world (Table 8) [29,[74][75][76][77][78][79][80][81][82][83][84][85][86], with details reported earlier [87].The variability of top drugs and drug classes across countries and populations (Table 9) is likely due to the variability of diseases and the various treatment options, perhaps also to differences of genetic predisposition.In all studies above, uniformity was achieved by using RUCAM globally and nationwide (Table 9) as the original RUCAM [17] and more recently the updated RUCAM [19], used, for instance, in a study in Japan on epidemiology and management of DILI with the note in the title on the importance of the updated RUCAM [91].Consequently, for future studies, the prospective use of the updated RUCAM is recommended [19] without combination with another tool, applying liver injury criteria of ALT or ALP thresholds (Figure 1) and the inclusion of cases with a probable or highly probable RUCAM grading only and with exclusion of possible cases.
The cases of iDILI shown above (Table 10) were assessed by the original RUCAM [17] or the updated RUCAM [19], showing RUCAM-based scores and causality gradings of probable or highly probable and ignoring cases with a possible or lower causality gradings.There is an appreciated trend to increasingly use the updated RUCAM with correct quotation and mention of RUCAM scores with causality gradings (Table 10) [60, 94,95,.
The updated RUCAM was used in at least 5/9 reports (Table 11) [60, 115,118,137,138].An analysis of previous reports on increased LTs in patients with COVID-19 revealed that many of them did not use RUCAM to confirm or exclude DILI, although the patients were heavily multi-medicated due to the severity of the clinical course and substantial comorbidity [144].However, various authors recommended the use of the original or updated RUCAM in future cases, which necessitates additional prospective studies using the updated RUCAM prospectively rather than retrospectively as done until now in virtually all published cases, because the issue of DILI was largely neglected despite the high number of multi-medicated COVID-19 patients [144,145].The existence of iDILI in a COVID-19 cohort will inevitably confound the clinical features of COVID-19 described so far because there was not yet any differentiation from each other considered in published cases.To make this point more clear, clinical features of COVID-19 are confounded and must be separated from the co-existing features of iDILI [144].The updated RUCAM [19] was used in 124 males and 36 female patients [137], providing in 82 patients a probable causality grading based on a RUCAM score of ≥6 and in 78 cases a possible causality ranking based on a RUCAM score of ≥3.The high number of cases with a possible causality grading could have been avoided by using a prospective and proactive study protocol.The mean number of drugs per patient was 14.7 (SD 7.6), whereby 98.1% received more than 5 drugs.Among the used drugs were acetaminophen, azithromycin, ceftriaxone, dexketoprofen, doxycycline, enoxaparin, hydroxychloroquine, interferon, levofloxacin, lopinavir, metamizole, omeprazole, pantoprazole, piperacillin/tazobactam, remdesivir, ritonavir, and tocilizumab.
Jothimani, 2021 [141] India (cases, n = 1) (drugs, n = 4) RUCAM was used without clear definition of the version applied [17,19] in a male COVID-19 patient [141], who suffered from DILI after using the oral anticoagulant dabigatran, for which a RUCAM score of 7 corresponding to a probable causality was provided.Additional medications included enoxaparin, esomeprazole, and methylprednisolone.It was outlined that the cause of liver injury is multifactorial in COVID-19.
Kumar, 2021 [142] India (cases, n = 3) (drugs, n = 3) In this study of three patients (2 females, one male) with COVID-19, each treated with favipiravir that caused DILI, RUCAM was used without specifying the RUCAM version applied [142].The updated RUCAM was likely used, which requires the exclusion of hepatitis E virus (HEV) infection [19], a parameter considered in the present study [142] that was not an element of the original RUCAM [17].For all three patients, a RUCAM score of 7 was found consistent with a probable causality level [142].Of note, the second patient also used acetaminophen, and the third patient was treated with entecavir for his hepatitis B-related cirrhosis, currently with absence of serum hepatitis B virus (HBV) DNA.
Deng, 2022 [118] China (cases, n = 2) (drugs 2) In 2 COVID-19 patients [55], the updated RUCAM was used [19], leading to a score of 8 for a probable causality for the male patient treated with ibuprofen and with a score of 9 for a highly probable causality for the female patient, who used acetaminophen [118].In

First Author Case Details of RUCAM-Based iDILI in COVID-19 Patients
Sigurdason, 2023 [143] Iceland (cases, n = 3) (drugs, n.a.) In a 2020 retrospective population-based study of Iceland on liver injury, 3/225 hospitalized patients with COVID-19 met the criteria of RUCAM-based DILI but achieved only a possible causality grading, leading to a conclusion of vague clinical features except that the affected population hospitalized in Iceland is not large enough to detect iDILI (between 1/10,000 and 20,000 inhabitants).
Table 13.DILI characteristics in COVID-19 patients, derived from a cohort of cases with causality assessment by RUCAM [144].

RUCAM-Based DILI Characteristics in COVID-19 Patients
• The male gender prevailed, and age was in a range of 45 and 57 years • Hepatocellular injury was more commonly observed than cholestatic or mixed injury • Multi-medication is likely a risk factor for liver injury • The list of used drugs is impressive but co-medication was evaluated smoothly • RUCAM can be well used in retrospective studies, although the better approach is using a prospective design • High RUCAM-based causality gradings can be well achieved in retrospective studies

• One study provided excellent external validation results of the updated RUCAM
• The large number of RUCAM-based iDILI cases can certainly foster the global use of RUCAM in addition to the 81,856 cases published so far • The high quality of RUCAM based iDILI cases cannot compensate the low quality of abundant published reports on COVID-19 patients with non-RUCAM based cases because these were, at best, only narratives describing the multiplicity of drugs patients were using Note: The excellent data on the validation of the updated RUCAM are briefly summarized in Table 6 and represent one of the highly appreciated external validations described in a report that explicitly calls for the determination of causality of DILI in patients with COVID-19 clinical syndrome [60].Abbreviations: COVID-19, coronavirus 2019; DILI, drug-induced liver injury; RUCAM, Roussel Uclaf Causality Assessment Method.
Mechanistic steps of drugs implicated in RUCAM-based DILI were proposed in several of the eight reports of cases with a verified diagnosis of DILI [60, 115,118,136,137,[140][141][142] and briefly summarized [144].DILI by tocilizumab (TCZ), a humanized recombinant monoclonal antibody that functions as an IL-6 receptor antagonist fighting against the cytokine storm, may be triggered by its binding to IL-6 receptors.Polypharmacy and pretreatment with potentially hepatotoxic drugs such as lopinavir/ritonavir could have facilitated the development of TCZ liver injury via molecular drug-drug interactions and drug-metabolizing enzyme induction/inhibition.In a large case series, hepatic steatosis was considered as a risk factor of liver injury.As this type of liver disease is closely related to being overweight and obesity, both of which commonly exert the induction of hepatic microsomal cytochrome P450 (CYP) 2E1, a role of this CYP isoform can be assumed.Another report focused on molecular details related to CYP 3A4, potently inhibited by ritonavir, which could promote liver injury by azithromycin through molecular mechanisms at the CYP level.Molecular mechanisms for enhancing DILI during inflammation could also be associated with ROS (reactive oxygen species) generated by inflammatory cells, possibly via myeloperoxidase, an enzyme located in inflammatory cells such as macrophages and neutrophils, in addition to immune mechanisms shown in a small subset of DILI cases.For liver injury by dabigatran, the molecular mechanism of the injury was assumed to be related to an idiosyncratic rather than intrinsic reaction.The DILI by favipiravir or its metabolites was ascribed at the molecular level primarily to an idiosyncratic reaction.However, continuous drug use causes self-inhibition of its hepatic metabolism, which may increase the favipiravir/inactive metabolite ratio, a possible contributory factor for the injury like a high drug intake.According to another proposal, a high loading dose of the drug and using other hepatotoxic drugs may represent the molecular and mechanistic basis of the liver injury.
Authors appreciated the broad use of RUCAM [60, 115,118,136,137,[140][141][142] and followed most of the instructions for its application [17][18][19]23,63] including the RUCAM worksheet for hepatocellular injury, cholestatic and mixed liver injury, as well as the RUCAM interpretation of ALT or ALP variations [63].Considered as an appeal, recent unpublished cohorts should be re-examined for RUCAM-based DILI cases because in the future, new cases cannot be expected due to the decline of the COVID-19 epidemic.

RUCAM and DILI Related to Immune Checkpoint Inhibitors
Major interest focused recently on RUCAM in connection with DILI related to immune checkpoint inhibitors (ICIs) [146][147][148], where this method was used, proposed, or discussed in many articles [54,111,112,[146][147][148][149][150][151][152][153][154], but not discussed in another report with a focus on RUCAM-based iDILI cases, although differentiation of iDILI by ICIs from autoimmune diseases such as genuine autoimmune hepatitis (AIH), primary biliary cholangitis (PBC), primary sclerosing cholangitis (PSC), and their overlap syndromes was proposed [15].ICIs were used to treat patients with malignancies, which may confound the diagnosis of iDILI in addition to multi-medication due to advanced age and major co-morbidities.These conditions call for a careful causality assessment using RUCAM [147], which was done, for instance, in a well-described case report from Israel published in 2022 with the title: acute liver failure following a single dose of atezolizumab, as assessed for causality using the updated RUCAM [111].In this report, a RUCAM-based score of seven was achieved corresponding to a probable causality level for atezolizumab as a cause of the acute liver failure, while acetaminophen as co-medicated drug received only a score of four in line with a possible causality grading [111], substantiating that the updated RUCAM can differentiate causality between co-medicated drugs [19].
The importance of using robust causality methods such as RUCAM was proposed in an article from the US showing that liver injury is most commonly due to hepatic metastases rather than DILI during pembrolizumab immunotherapy [54].Since the opinions of experts were not widely available, the standardized scores of the original RUCAM [17,18] were calculated for each of the 70 cases.However, only 8/70 cases achieved a RUCAM score >8 (highly probable) since most patients did not undergo testing for competing causes of the liver injury and then received a lower score [54].In 50/70 cases, there was a possible or unlikely RUCAM-based causality, and among these, the most identified alternative cause of the liver injury was progressive hepatic tumor metastases (56%), while other etiologies included malignant biliary obstruction (4%), non-hepatic disease (9%), and other biliary obstruction or unknown cause (2%) [54].Due to additional limitations of the study like the retrospective study protocol, data extraction from a so-called self-serve electronic medical record (EMR) search tool, the presented results remain vague with a critical note [54] that prior publications on this topic often have simply described as narratives on the incidence of liver injury during treatment but not adjudicated for the cause [155,156], which was professionally analyzed in medical malpractice liability disputes on iDILI calling for the use of the updated RUCAM [155].In another narrative report from China published in 2023, the suitability of the updated RUCAM for the iDILI diagnosis related to ICI treatment was unreasonably discussed ignoring actual publications [146].In this narrative publication, many statements were derived from the literature were purely speculative and not based on evidence by RUCAM.As opposed and in view of alternative causes in patients treated with ICIs, it is strongly recommended that new cases are prospectively assessed by the updated RUCAM [147,151] to shed more light in this special liver injury to be included in review articles and clinical guidelines [147].Based on studies that used RUCAM, insights are provided on iDILI details caused by ICIs, with the call that it is of utmost importance to use a formal scoring system such as RUCAM to assess risk factors, alternative causes, and response to cessation and reexposure of a given drug [150].Otherwise, clinical characterization of iDILI by ICIs cannot be well described in studies [157][158][159][160][161][162] and would be acceptable only if concurrent alternative causes were carefully ruled out.

RUCAM in Search for Alternative Causes
Clinical features of iDILI have no specificity and are often confounded by alternative causes [1,19,158,163,164] as shown quantitatively after an analysis of 4555/13,335 cases (34.2%) using RUCAM [164].In this study, alternative causes included from top to down biliary diseases (11.89%), autoimmune hepatitis, tumor, ischemic hepatitis, systemic sepsis, hepatitis, liver injury by other comedication, virus hepatitis, past liver transplantation, alcoholic liver disease, fatty liver, nonalcoholic steatohepatitis, hepatitis C, cardiac hepatopathy, thyroidal hepatopathy, primary biliary cholangitis, Gilbert syndrome, cytomegalovirus infection, Epstein-Barr virus infection, hemochromatosis, Wilson disease, paracetamol overdose, postictal state, osseous disease, lymphoma, pre-existing cirrhosis, hepatitis B, benign recurrent intrahepatic cholestasis, rhabdomyolysis, polymyositis, Chlamydial infection, and human immunodeficiency virus infection (0.31%) [164].With this back-up information, among a cohort with iDILI patients there may be cases that were not DILI but rather due to non-drugs as causes.Consequently, in each suspected iDILI case, exclusion of alternative causes is mandatory [19], a reason why a checklist may be helpful whereby in addition to the published recommendations, a checklist can be of assistance (Figure 2) [19,164].
A discussion is warranted on HEV infections representing an alternative diagnosis in 5.78% of iDILI cohorts [164], although mandatory exclusion of HEV was already proposed in earlier reports published in 2007 [165], 2008 [166], and 2009 [167].The problem of missed HEV diagnosis in DILI cohorts was subsequently also seen by other groups [168][169][170][171][172][173][174][175][176].Historically, the virus could not be cloned for many years [177] until 1991 when a fulllength HEV genome was sequenced, and a diagnostic tool based on enzyme immunoassay was established but was not yet commercially available [178].Serological analysis of HEV infection remained problematic for many years because development of valid tests was retarded [177].This is why tests to exclude HEV infection were not included in the algorithm of the original RUCAM published in 1993 [17].With the test availability and application, HEV infections were detected in iDILI cases confounding the DILI diagnoses as they were not anymore DILI [168][169][170][171][172][173][174][175][176], with reminders to be aware of HEV in patients with suspected DILI [174], the unmet needs of HEV diagnosis in suspected DILI [175], and screening of HEV infection should be included in the current diagnostic scale for DILI in Japan [176].In 2016, the updated RUCAM included in its algorithm-appropriate tests to search for HEV infection [19].This is one of many reasons why the updated RUCAM version should be used in upcoming suspected DILI cases as done already in current cases (Table 11).Flow chart with checklist of differential diagnoses in cases of suspected iDILI.This compilation is adapted and derived from a previous open-access publication, and although not comprehensive, it is to be used as a guide and in connection with RUCAM [19].Abbreviations: AAA, anti-actin antibodies; AMA, antimitochondrial antibodies; ANA, antinuclear antibodies; ASGPR, asialo-glycoprotein-receptor; BMI, body mass index; CT, computed tomography; CYP, cytochrome P450; DILI, drug-induced liver injury; DPH, pyruvate dehydrogenase; HAV, hepatitis A virus; HBc, hepatitis B core; HBV, hepatitis B virus; HCV, hepatitis C virus; HEV, hepatitis E virus; HIV; human immunodeficiency virus; LKM, liver kidney microsomes; LP, liver-pancreas antigen; LSP, liver-specific protein; MRC, magnetic resonance cholangiography; MRT, magnetic resonance tomography; and p-ANCA, perinuclear antineutrophil.

RUCAM in Search for Pharmacotherapy Options of iDILI
RUCAM helped search for pharmacotherapy options of iDILI including cases of a subtype characterized by serological autoimmune features and diagnosed as druginduced autoimmune hepatitis (DIAIH), for which corticosteroids are the first choice of treatment as shown in a study that applied the updated RUCAM to exclude competing causes [179], perfectly done by also using a specific diagnostic algorithm of genuine autoimmune hepatitis (AIH) [180] in a separate control cohort of AIH [179].Differentiation of DIAIH from the usual DILI is commonly well achieved [179][180][181][182], even if the two diagnostic algorithms of iDILI and AIH were used in combination for DIAIH [181,182], but under these conditions, preference should be given to the updated RUCAM [181] rather than the original RUCAM [182].RUCAM-based studies showed DIAIH causally related to drugs [183] like antimicrobials [184,185], atorvastatin [183], augmentin [184], ceftriaxone [185], diclofenac [186], direct oral anticoagulants [179], hydralazine [186], irbesartan [187], infliximab [188,189], isoniazid [186], ketoprofen [184], methyldopa [186], minocycline [186], nimesulide [184], nitrofurantoin [186,188,190,191], non-steroidal antiinflammatory drugs [179,184,190,192], sorafenib [183], and statins [179,185,191].Direct evidence for the involvement of the innate immune system in the DIAIH group was convincingly shown with causative drugs such as diclofenac, indomethacin, levofloxacin, and phencoumon by studies of monocyte-derived hepatocyte-like cells in iDILI cases assessed by the updated RUCAM [193].These results support the contention that monocytes are part of the innate immune system [147,185,[194][195][196] with the potential of an additional pharmacotherapy option for this cellular-driven liver injury [147].Direct evidence for an involvement of the immune system in idiosyncratic DILI was also provided by its rare association with the immune-triggered Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) caused by a small group of drugs [195].Causality of idiosyncratic DILI was evaluated by RUCAM and of SJS/TEN using the Algorithm for Drug Causality for Epidermal Necrolysis, which was highly probable or probable in all cases.In the context of immunology, among the drugs implicated in triggering iDILI, 58.3% are metabolized by CYP isoforms [197,198], and four patients with RUCAM-based iDILI caused by drugs metabolized by CYP isoforms like the volatile anesthetic sevoflurane, showed serum antibodies against cytochrome P450 (CYP), whereby RUCAM achieved causality gradings of highly probable in all cases [199].These results were confirmed by a subsequent study of sevoflurane and desflurane, another modern volatile anesthetic, in which sevoflurane was applied mostly alone [61].Three patients with serum anti-CYP 2E1 antibodies and iDILI reached a RUCAM score of 12, 7, and 6.The association of serum anti-CYP isoforms with iDILI by sevoflurane implicates an immunological involvement in this process but not necessarily a causal immune association leading to the liver injury.
For iDILI unrelated to autoimmunity, valid data recommending pharmacotherapy options cases induced by hundreds of drugs are limited because it requires randomized controlled trials (RCTs) using RUCAM-based iDILI cases to verify treatment efficiency [147].A recent analysis revealed that only eight RCTs have been published, and only two of them were DILI cases evaluated for causality using RUCAM [200], representing a significant methodology flaw, as results of DILI RCTs lacking RUCAM were often misleading since they were confounded by competing non-drug causes.In another review article from China, pharmacotherapies for iDILI were suggested after analysis of mostly narrative DILI studies lacking case evaluation by RUCAM [201]; although, the Chinese Drug-induced Liver Injury (DILI) Study Group, the Chinese Society of Hepatology (CSH), and the Chinese Medical Association (CMA) strongly recommend in their CSH guidelines for the diagnosis and treatment of drug-induced liver injury the use of a valid diagnostic algorithm such as the updated RUCAM to verify causality [24].
A positive therapeutic effect was described in a study from China that perfectly used the updated RUCAM with scores ≥6, pre-defined thresholds of ALT and ALP, and exclusion of competing causes in liver injury cases with patients, who were treated with corticosteroids plus glycyrrhizin [202].In this single-center randomized open-label trial, cases of DILI and HILI cases were included with separate evaluation of the two cohorts and their subgroups in line with CSH guidelines, recommending the use of the updated RUCAM in future trials with more patients, with some open questions that were clarified subsequently [203,204].This prospective study from China also attracted the interest of another group working on the same topic but used a less straightforward approach [205] compared to the Chinese study [202].
Presented as an international, multi-center, propensity score-matched evaluation with a focus on benefits and risks of corticosteroids in DILI, major shortcomings of the so-called bona fide analysis were evident [205]: (1) the cohort lacked homogeneity as DILI cases with or without immunological features were included in a mixed fashion, confounding the results as not restricted to non-immune DILI cases, because 44% of patients had positive autoantibody titers and 24% were diagnosed as drug-induced autoimmune-like hepatitis syn DIAIH using debatable criteria; (2) lacking an appropriate prospective protocol, data had to be assessed retrospectively with cases already available in two different DILI registries, one in Spain and the other one in the US; (3) rather than using the updated RUCAM of 2016, the Spanish cases were evaluated partially considering RUCAM criteria derived from various publications dating back to 1990 or 2011 without considering HEV as a mandatory element to be assessed but with the inclusion of bilirubin, a common parameter of the widespread genetic Gilbert syndrome, and cases with a possible causality grading; (4) the US cases were evaluated with the network methodology, which lacks robust internal and external validation as based on global introspection of experts with different subjective opinions and the inclusion of cases with a possible causality grading; (5) causatives were not explicitly defined but may be included according to introductory and text statements in addition to conventional medications, herbal products, or so-called dietary supplements that cause HILI but not DILI; (6) due to the retrospective study protocol, details of corticosteroid treatment like timing, route of administration, and dosing were not pre-defined but relied on the discretion of the attending physician; (7) in only 30% of cases, details of corticosteroid treatment were available, indicating poor quality of prospective case data acquisition and documentation; (8) in 106/724 cases (14.6%), treatment included corticosteroids, reflecting previous suggestions to treat on a case-by-case decision; (9) remarkable and debatable is the statement that both internal and external validity of the findings is ensured; (10) any assumed beneficial effect may be caused by DILI cases with immune features that are known responders to corticosteroid treatment; and (11) in sum, evidence is lacking that patients with non-immune DILI may benefit from glucocorticoids but study protocol precluded conducting subgroup analysis.

RUCAM in Genetic iDILI
The occurrence of iDILI in a few individuals exposed to recommended doses of a drug suggests a role of genetic variation responsible for the injury [148,206].In 2009 and after analysis of iDILI cases, which achieved a probable or highly probable RUCAM causality grading in 92% of the cases, direct evidence was presented that iDILI is triggered partly by genetic susceptibility of human leucocyte antigen (HLA) alleles [207].In this report alongside a genome-wide association study (GWAS), HLA-B*5701 genotype was noted as a major determinant of iDILI due to flucloxacillin.Comparable results of HLA genotypes were found in other RUCAM-based iDILI cases for several drugs including antituberculosis drugs [208], nitrofurantoin [45,209], amoxicillin-clavulanate [78,79], diclofenac, azathioprine, isoniazid, fenofibrate [45], and flucloxacillin [51,210].However, there was a lack of data reproducibility with respect to cases of iDILI due to amoxicillin as evaluated by RUCAM [211] as well as regarding nitrofurantoin considering iDILI cases evaluated for causality using global introspection, a non-RUCAM approach lacking proper validation, and individual element scoring for appropriate objective quantitative results [212].An assumed HLA association of liver injury by amoxicillin-clavulanate has been proposed already in 1999, but results remained vague because cases were assessed for the liver injury pattern only but not for causality using the original RUCAM of 1993 [213].Despite many studies, current HLA data are only a little step forward to partially characterize iDILI [148], and as genetic markers, they exhibit a high negative predictive value and low positive predictive value, limiting their clinical use to prospectively predict iDILI risk [214].The overall clinical value of HLA B*5701 pre-assessment in an individual patient, for whom a treatment with flucloxacillin is planned, is in question since there is less than 1/500 chance that the patient will develop iDILI in the case of HLA B*5701 positivity [206].Although HLA studies showed an association of genetics with iDILI caused by a limited number of drugs, their contribution in elucidating additional mechanistic details in IDILI remains marginal [209], let alone its value as a diagnostic biomarker or causality tool.The question is why several drugs causing iDILI such as isoniazid, diclofenac, azathioprine with other thiopurines, ciprofloxacin with other fluoroquinolones, atorvastatin with other statins, nimesulide, interferon beta, and fasiglifam were without detectable HLA association, and why many published HLA cases were not assessed by RUCAM to verify the iDILI diagnosis [215].In essence, the early recognition was perfect to apply RUCAM for valid causality evaluation in HLA studies [207] and to be viewed as a general recommendation for other studies on this topic.Prospects in HLA genetics were outlined [209,215] but must include iDILI cases evaluated by using the updated RUCAM with a high causality grading of probable or better, highly probable.

RUCAM, iDILI, and the Microbiome Dysbiosis
The role of microbiome dysbiosis in causing iDILI is best studied in RUCAM-based iDILI cases [135,187,216].For instance, a prospective study on the urine microbiome in patients with iDILI caused by anti-tuberculosis drugs revealed a microbiome risk for the liver injury evaluated prior to treatment initiation [135].Included in the study were patients with a causality grading of ≥6 based on the updated RUCAM, viewed as the recognized standard for the diagnosis of iDILI.The study was done using urinary microorganisms, which are considered to reflect the microbiota of the entire body including the intestinal tract, oral system, and respiratory system.Upregulated urinary Negativicoccus and Actinotignum at the start of drug treatment were identified as risk factors and predict a severe course of iDILI by anti-tuberculosis drugs [135].These findings require verification with more patients but are interesting, although currently confined to patients with tuberculosis raising the question of whether similar data can be achieved in patients receiving drugs other than tuberculosis drugs.

RUCAM and Metabolomics in iDILI
There are several reports on RUCAM-based iDILI studies with a focus on the metabolomics technology [135,[217][218][219] defined as the analysis of metabolites in a biological sample like blood or urine, which can help characterize metabolic alterations that underly the disease and search for potential biomarkers to be used to either diagnose the disease or monitor efficiency of treatments [220].Metabolomics belongs to the omics group and differs from genomics that focuses on the structure, function, evolution, mapping and editing of genomes found in an organism [221][222][223], and from proteomics that covers the section of proteomes, a set of proteins in an organ or organism [224][225][226].
As a first example of the use of metabolomics in iDILI, in a prospective study from China [135] assessed by the updated RUCAM [19] in line with a recommendation in the CSH guidelines for the diagnosis and treatment of DILI [24], the mechanism of the iDILI was explored [135] by urine metabolomic analysis in a cohort of patients with anti-tuberculous drug treatment and iDILI of probable and highly probable causality.A total of 28 major metabolites were screened out, involving bile secretion, nicotinate and nicotinamide metabolism, tryptophan metabolism, and ABC (ATP-binding cassette) transporters prior to anti-tuberculosis drug therapy [135].The analysis showed significant differences of the metabolic profile between the DILI group and the non-DILI group that may help predict the risk of liver injury after start of the treatment with anti-tuberculosis drugs.
The second example refers to a Spanish metabolomic analysis of patients with DILI assessed by the original RUCAM, which provided scores of >6 (probable and higher causality grading) [217].Although the use of RUCAM is perfect, the rationale of the study remained theoretical and vague despite lengthy introductory critical remarks on the current liver injury pattern and attempts to improve DILI phenotyping.As expected, metabolomic results of the hepatocellular injury are different compared with those of the cholestatic liver injury regarding serum bile acid profiles including primary, secondary, conjugated, and non-conjugated bile acids [217].However, these data were not exciting and add little to current differentiation by LTs easily obtained in a clinical setting (Figure 1) [19] as opposed to cost-effective metabolomic analyses performed on an ultraperformance liquid chromatograph (UPLC) [217].Due to fraudulent data on diagnostic biomarkers including high mobility group box 1 (HMGB1) initially proposed and later retracted by the EMA and the FDA [66], the discussion on this topic now seems irrelevant [217].
The third example of a metabolic study on RUCAM-based iDILI cases from China identified and quantified 24 types of bile acids that can predict the severity of iDILI [218].Rather than using the updated RUCAM [19] as recommended by the CSH guidelines [24], the study applied the original RUCAM [17,18], which provided causality gradings for 161 cases of highly probable in 5 cases, probable in 109 cases, and possible in 47 cases [218].Because possible cases were included in the analysis, resulting data must be seen cautiously.
The fourth example refers to a Chinese metabolomic study, which used the original RUCAM for the iDILI cases, providing causality scores of six and more that include probable and highly probable causality gradings [219].The study focused on plasma metabolomic and lipidomic alterations associated with anti-tuberculosis iDILI.
Finally, many narratives on the metabolomics issue in iDILI with methodology shortcomings were published but prevent any further discussion or critical analysis of the results [227][228][229].For instance, in one study from China, the RUCAM version used was unclear [227], in another report from Spain, RUCAM use was discussed but not applied [228], and in a third analysis from China, the use of RUCAM remained vague [229].For reasons of completeness, the metabolomic technique was also applied by in vitro studies to assess the potential of liver injury by new drugs [230].Published results are rare, and the significance of metabolomic analysis done in vitro largely dependent on the characteristics of the different hepatic cell types used and their specific requirements.

RUCAM-Based iDILI Features
The classical features of liver injury pattern syn phenotype like hepatocellular injury, cholestatic, or mixed liver injury identified alongside the consensus meetings for the original RUCAM [17,18] were used up to now in around 100,000 DILI cases, occasionally without giving credit to the initiators, including the 81,856 RUCAM-based iDILI cases published until mid-2020 [20].A systematic review from Australia with meta-analysis according to PRISMA guidelines (Preferred Reporting Items for Systematic reviews and Meta-Analyses) compared hepatocellular and cholestatic patterns of drug-induced liver injury and presented a list of 12 studies comprising 4290 patients derived from national and single center registries and multi-center registries [231].Ten studies used the international consensus criteria of RUCAM while two studies applied a structured expert opinion process for their causality assessment.Patients with cholestatic DILI demonstrated similar rates of acute liver failure (ALF), and liver-related deaths compared to patients with hepatocellular DILI, and patients with cholestatic DILI were significantly more likely to experience chronicity compared to patients with hepatocellular DILI.Of note, the so-called Hy's law points out that DILI with jaundice is associated with high mortality when the liver injury is of the hepatocellular type, whereas the current analysis showed that cholestatic injury can also be associated with similar rates of ALF and liver-related fatality, as well as higher rates of chronic DILI [231].This note was accompanied by the conclusion that Hy's law as a potential prognostic marker for iDILI requires validation.

Epidemiology of iDILI and RUCAM Use
The note from Japan that RUCAM is used in all epidemiology reports on DILI internationally was stimulating [91].Indeed, the most important epidemiology studies used the international consensus criteria of RUCAM [234]: the rate of DILI was 13.9 per 100,000 persons/year in France [235], 2.4 in Spain [236] and with 3.42 broadly confirmed again in Spain [74], 2.3 in Sweden [237], 3.7 in Iceland [80], and erroneously reported as 12.0 in Korea because only a small percentage comprised prescription drugs but included herbal products and dietary supplements outnumbered the drugs by far [238].

Pharmaceutical Firms
Authors as employees or external advisors of pharmaceutical firms successfully applied RUCAM for the causality assessment of suspected iDILI in many clinical trials or after the marketing of new drugs to provide transparency and clarity whether the drug under consideration injured the liver [19,20].A good example of efficacy in long-term trials was the study of iDILI caused by ximelagatran, where the use of RUCAM was appreciated and helped the pharmaceutical firm remove the drug from the market [26].This approach was straight forward, and there is no current need to re-invent the wheel by creating a new tool for causality assessments as discussed in a recent article [239].Some co-authors who previously engaged in publications using RUCAM [74,75] now prefer an unvalidated method based on global inspection and mere opinions not based on evidence and reproducible data.This approach certainly provides disputable results as opposed to objective and validated causality assessment methods like the updated RUCAM mentioned in this manuscript but not recommended for mandatory own use [239].
Remarkable are additional points that warrant critical analysis of the overview on causality assessment for DILI in clinical trials [239]: (1) surprisingly, seeking an expert opinion to diagnose DILI was considered as the 'gold standard' approach, a statement not based on evidence; its use by pharmaceutical companies and regulatory agencies as claimed by the authors remains debatable; (2) challenges were seen by the authors in both clinical trials and post-marketing settings that include missing data and collection of an appropriate minimal dataset to enable adequate assessment of causality for DILI; (3) firms see specific difficulties of missing included immunology data, diagnostic imaging, hepatitis serology, liver synthetic parameters, details on confounders, and complete medical history, conditions not allowing for a good quality causality assessment; (4) authors proposed the development of a minimum data set to balance considerations of feasibility and completeness and to ensure a more complete set of data to evaluate cases but there was the note that such suggestions have not been widely adopted, although experts in the field recommend such a list of minimal clinical and laboratory data considered essential for DILI diagnosis; in fact, there was no unified approach for data collection across the industry even within clinical development, leaving this issue unresolved; (5) authors developed their report, as they said, in collaboration with academia and government subject matter experts and the IQ DILI Causality Working Group (CWG) and passing several face-to-face meetings, but overall resulting proposals were disappointing despite much effort and remained at a low quality level; (6) another aim of the authors was to promote a structural and universal approach for DILI causality assessment, but nothing substantial and promotional was found in the text, likely because most of the interviewed firms reported that they will not be able to make data available for validating new methodologies due to the complexities of having patient consents, having finalized the available data, and having anonymized patient-level data; (7) the authors' statement that there is a need for a uniform approach is correct because the 2009 US Food and Drug Administration (FDA) did not provide necessary details as standards for how causality assessments should be conducted, but no uniform approach was provided; (8) authors classified the US network method as a scoring system, but actually, this system provides only a subjective final scoring based on a percentage range not considering individual element scoring; (9) authors claimed that the expert approach is believed to be the most effective approach to DILI assessment and should be the primary method used, which is a vague proposal far away from the evidence and reality; (10) authors believe that subjective expert opinion is accepted practice (standard) superior to standardized instruments, which is certainly again debatable, also in view of their note that RUCAM is well published and objective; (11) authors mentioned the eDISH (Evaluation of Drug-Induced Serious Hepatotoxicity), which is not a CAM; (12) as opposed to authors' statement, RUCAM showed good internal observer reproducibility and received external validation by interrater reliability in three studies mentioned in Table 4 above; (13) contrary to authors' view, RUCAM is validated for suspected DILI cases from all sources, an extra validation of cases from trials is not needed; (14) disturbing is the discussion on biomarkers by the authors mentioning that a subset of these has recently received regulatory support from both the EMA and the FDA for more systematic use in an exploratory development setting; whereas, the quotation of the respective reference was correct, expert authors forgot to mention that, due to fraudulently published results from European scientists, both agencies officially retracted their initial support long ago [66]; (15) the majority of companies were said to be using LiverTox to identify alternative causes, a strange statement when looking on its debated qualification [64,240]; and ( 16) the methodology and assumption flaws of causality assessment for DILI in clinical trials certainly require re-consideration.

Regulatory Agencies
Because pharmaceutical firms continue mere discussions without decisions about the best method of causality for their cases of suspected iDILI [239], this represents a particular challenge for regulatory agencies.To overcome these problems, national regulatory agencies are well advised to call for iDILI to be assessed using the updated RUCAM.

DILI Registries
DILI registries have problems with data collection, interpretation, and publication [64], requiring adherence to the recommendations of the updated RUCAM.This would enhance trust in the valuable work of the registries.

LiverTox
The poor case data quality comprised in the US LiverTox database [241,242] remains a problem [64,240], only to be solved by including cases assessed by the updated RUCAM.LiverTox scientists promised the use of RUCAM for many years but did not get it done, now considered as the LiverTox paradox [240].Even worse, drugs comprised in the LiverTox database were arbitrarily divided into different categories of likelihood for causing liver injury based on the number of reports in the literature without any formal causality assessment [241,242]: category A, highly probable causality (>50 published cases), category B, highly likely (12-49 published cases), category C, probable (4-11 published cases), category D, possible (1-3 published cases), and category E (no published cases) [241][242][243].In other words, the higher the number of published cases on a drug, the higher the causality level, which is a debatable approach.A separate analysis of the highly probable category A revealed in 2018 that among the listed 48 drugs, no RUCAM-based case was identified for 9 drugs (24%) [87]: busulfan, dandrolene, didanosine, efavirenz, floxuridine, nevirapine, quinidine, sulfonamides, and telithromycin.Gaining trust in LiverTox case data quality will require substantial efforts.

Guidelines
Compared with the excellent CSH guidelines on diagnosis and treatment of iDILI from China recommending among others the mandatory use of the updated RUCAM [24], abundant DILI guidelines from other countries were published, often with contradictory recommendations and missing proposals to use the updated RUCAM.A uniform guideline together with the one of the CSH is encouraged to provide consistency.

Getting All under One Hat
An editorial of 2015 in Gastroenterology on iDILI called for expanding our knowledge by enlarging population analysis with prospective and scoring causality assessments [244].This aim was well achieved in the meantime: first, the global use of RUCAM preferentially based on a prospective study protocol to obtain high causality gradings, and second, by providing the updated RUCAM as a user-friendly diagnostic algorithm with elements individually scored for robust causality assessment.To increase standardization, comparability, transparency, and consistency in data collection, analysis, and specific clinical and pathogenetic presentation, researchers and pharmaceutical firms interested in iDILI issues should take place under the same hat, together with the updated RUCAM as the causality tool in review articles and clinical guidelines for the diagnosis and treatment of iDILI.Much is already achieved but a little more effort is still needed to get it all under one hat, as shown, for instance, in a historical picture from Germany trying to get restless bourgeois of a town under one hat (Figure 3).

Conclusions
Idiosyncratic drug-induced liver injury presents as a complex and multifaceted entity with features characterized by variability.RUCAM helps assess causality for suspected drugs with variable chemical structures causing liver injury.The global use and appreciation of the original and updated RUCAM is fascinating and will encourage scientists, physicians, regulatory agencies, DILI registries, and pharmaceutical firms to add more DILI cases to the current list of 81,856 cases published until mid-2020, now preferentially using the updated RUCAM.Having iDILI and the updated RUCAM under one hat as a strong harmonizing concept will help further define iDILI characteristics among various populations offering a unique approach that includes a standardized causality assessment method resulting in robust and consistent data.

Figure 1 .
Figure 1.Classification of the liver injury pattern.

Figure 3 .
Figure 3. Historical picture from Germany.Kindly provided with allowance to be published by Dieter Ante, M.D., Germany.

Table 1 .
[20]ranking countries providing DILI cases assessed for causality by RUCAM.Data are derived from an open-access article[20].

Table 2 .
Compilation of 20,311 RUCAM-based DILI cases as published by authors of the United States; data were derived from a previous open-access analysis

Table 3 .
Experiences and weaknesses of the US DILI network method, as published

Table 6 .
Challenges and limitations of RUCAM.

Table 8 .
Top list of worldwide drugs causing DILI assessed for causality with RUCAM.

Table 9 .
National top-ranking drugs and drug classes implicated in RUCAM-based iDILI compared with global data.

Table 10 .
List of known drugs causing iDILI, assessed for causality by RUCAM

Table 11 .
New drugs causing iDILI, assessed for causality by the updated RUCAM.

Table 12 .
[136]8] of RUCAM-based iDILI in COVID-19 patients.Reported is a case of DILI by tocilizumab (TCZ) in a male patient with COVID-19 infection that caused a cytokine storm[140].Using the original RUCAM[17,18]instead of the commonly preferred updated RUCAM[19], causality for TCZ was probable based on a RUCAM score of 8.Such high causality grading is commonly achieved with complete data sets collected prospectively during the clinical course.TCZ is a humanized recombinant monoclonal antibody that acts as an IL-6 receptor antagonist by specific binding to IL-6 receptors.Preexisting liver disease was excluded as well as anoxia leading to liver hypoxia.It was noted that slightly elevated transaminases were detected before TCZ administration.Comedication included azithromycin, ceftriaxone, chloroquine, lopinavir, methylprednisolone, and ritonavir, but none of these drugs were considered causative for the liver injury, although a contributory role of the previously used antiviral drugs (lopinavir/ritonavir) is possible.This study analyzed 830 COVID-19 cases with liver injury.This is the largest study cohort evaluated for causality[136], using the updated RUCAM [17].Among 74/830 cases, the RUCAM score was >3, corresponding to a possible, probable, or highly probable causality.To achieve a homogeneous cohort, a good approach would have been to include only cases with a probable or highly probable causality.The concomitant drugs abidol, acetaminophen, oseltamivir, and ribavirin were discussed.For this retrospective study, all data were retrieved from the digital medical records during hospitalization.Important note is when multiple drugs in combination are used in COVID-19 patients, the RUCAM score is required to evaluate the risk of DILI of each drug.