20
Diagnosing Drug‐Induced Liver Injury

Guruprasad P. Aithal^1,2

¹Nottingham University Hospitals NHS Trust, Nottingham, UK

²University of Nottingham, Nottingham, UK

KEY POINTS

The diagnosis of drug‐induced liver injury (DILI) remains a challenge in many cases: the risk of unnecessary discontinuation must be balanced against the risks of continuing treatment.
DILI may be caused by prescribed medications, over‐the‐counter medications, herbal remedies, and natural remedies, so a full history should be obtained.
DILI should always be considered in cases of unexplained abnormal liver tests.
While some drugs have a characteristic pattern of liver damage, the spectrum associated with any one drug may be broad.
Causality assessment tools such as the Roussel Uclaf Causality Assessment Method (RUCAM) may help in attributing causality to a drug.
The diagnosis of DILI includes one of the three criteria: alanine transaminase elevation fivefold or more above the upper limit of normal (ULN), alkaline phosphatase twofold or more above ULN or alanine amino transferase (ALT) threefold or more above ULN with elevation of bilirubin twofold or more above ULN.
Evaluation should include a search for evidence of hypersensitivity (such as fever, rash, and peripheral blood eosinophilia); liver histology and genetic tests may also aid diagnosis.
Before making a diagnosis of DILI, other causes for liver damage should be considered.

Acute onset of liver injury attributable to the exposure to a drug taken in its therapeutic dose is referred to as . The term “drug” includes non‐prescription medicinal products, herbal remedies, and dietary supplements unless specified. Overdoses, whether intentional or inadvertent, are not discussed in this chapter. Distinction between intrinsic and idiosyncratic is hypothetical, as a subtle dose relationship (within the therapeutic range) exists in relation to DILI secondary to a number of drugs.

Challenges in the Diagnostic Approach

Identifying a particular drug as the etiology of an acute liver injury early is of critical importance, as its prompt withdrawal can reduce the risk of serious consequences. On the other hand, interruption of antituberculosis, antiepileptic, or antineoplastic agents may reduce the effectiveness of the treatment regimen. The general approach to diagnosis in clinical practice, as with other conditions, involves pattern recognition, assessing the pretest probabilities, and performing suitable tests to support or refute the diagnosis. However, confident diagnosis of DILI has remained a challenge because of a number of factors. First, over 650 medicinal products have been associated with adverse liver reactions; medications that accounted for DILI in consecutive patients enrolled into a cohort study are listed in Table 20.1. Polypharmacy has risen with the prevalence of multimorbidity. Truly distinct “signature patterns” of liver injury attributable to a particular drug or group of drugs are exceptions than a rule. Although lack of specific tests is highlighted as a barrier, specialist tests such as liver biopsy and genetic tests are underused.

We make rapid decisions, deemed necessary while managing medical emergencies or when evaluating many patients in a time‐limited manner. Our approach is aided by “heuristics,” strategies that provide short cuts to quick decisions, which in turn are pitfalls in diagnosis. In a large European study involving over 10 000 patients presenting with jaundice, doctors had an overall diagnostic accuracy of 77% at the time of initial assessment compared with the final diagnosis. However, diagnostic accuracy was 49% for DILI and it was ranked 15th of 18 categories of diagnoses considered in the study. Systematic evaluation would be able to overcome some of these limitations.

Be Aware of the Base Rates

Knowledge of how frequently particular condition occurs in different clinical contexts informs the pretest probability of a diagnosis. Considering DILI as a differential diagnosis in all acute presentation of liver injury is the first step in the systematic evaluation.

Raised Liver Enzymes

An international expert working group established the threshold for defining DILI, which includes any one of the three criteria: alanine transaminase elevation fivefold or more above the ULN, alkaline phosphatase twofold or more above ULN or ALT threefold or more above ULN with elevation of bilirubin twofold or more above ULN [1]. Aspartate amino transaminase (AST) is a substituted for ALT when the latter is not a part of the available panel of tests. Unexplained elevation of ALT five or more times ULN unrelated to the drug exposure have been observed in 0.3–0.4% of large population monitored regularly over one year. In a recent study involving 3155 patients on standard regimen of antituberculosis treatment, 218 (6.9%) developed ALT elevation between three and five times ULN [2]; of these, 193/218 (88.5%) resolved without significant (more than seven days) interruption in the treatment, while none of the 25/218 (11.5%) who developed DILI (ALT ≥ fivefold ULN) in this group progressed to acute liver failure. Lower thresholds of ALT elevation may be suitable for monitoring adverse liver reactions in clinical trials, where knowledge about the efficacy and safety of a medication is yet to be confirmed. However, unexplained elevation of ALT between above threefold ULN has been observed in 0.26% of people per year in the placebo group of a clinical trial and most were not explained by diagnosable liver disease. However, six of these who had raised bilirubin greater than twofold ULN all had liver disease requiring hospitalization. In a group of patients with atrial fibrillation attending routine outpatient care, 0.2% had new persistent elevation of ALT greater than twofold ALT. Cohort studies describing etiology underlying ALT/AST elevation of greater than 10‐fold ULN or greater than 20‐fold ULN have found 9% of these are due to DILI [3, 4].

Table 20.1 Etiological agents of consecutive 78 cases of drug‐induced liver injury (DILI) presenting to secondary care hospitals in the UK.

Drug group	Drug	Cases (n)	Total DILI (%)	Pattern of DILI
Antimicrobial	Co‐amoxiclav	13	16.46	Eight hepatocellular; three mixed; two cholestatic
	Flucloxacillin	12	15.19	Five hepatocellular; five mixed; two cholestatic
	Nitrofurantoin	4	5.06	Three hepatocellular; one cholestatic
	Doxycycline	2	2.53	Two cholestatic
	Pivmecillinam	1	1.27	Hepatocellular
	Isoniazid	1	1.27	Hepatocellular
	Total	33	41.77
Immune checkpoint inhibitors	Ipilimumab/nivolumab	11	13.92	All hepatocellular
	Nivolumab	1	1.27	Mixed
	Total	12	15.19
Statin	Atorvastatin	9	11.39	Four hepatocellular; five cholestatic
Antimetabolites	Azathioprine	3	3.80	Two mixed; one cholestatic
	Methotrexate	2	2.53	Hepatocellular
	Gemcitabine	1	1.27	Cholestatic
	Total	6	7.59
Antineoplastic	Asparaginase	2	2.53	One hepatocellular; one mixed
	Temozolomide	1	1.27	Hepatocellular
	Docetaxel	1	1.27	Mixed
	Total	4	5.06
Biologics	Infliximab	2	2.53	Hepatocellular
	Vedolizumab	1	1.27	Hepatocellular
	Total	3	3.80
Tyrosine kinase inhibitor	Zanubrutinib	1	1.27	Hepatocellular
	Panzopanib	1	1.27	Hepatocellular
	Total	2	2.53	Hepatocellular
NSAIDS	Ibuprofen	1	1.27	Hepatocellular
	Naproxen	1	1.27	Hepatocellular
	Total	2	2.53
Angiotensin ii receptor antagonists	Irbesartan	1	1.27	Cholestatic
Anesthetics	Propofol	1	1.27	Hepatocellular
Anabolic steroids	SD matrix	1	1.27	Cholestatic
	DNP (2,4‐dinitrophenol)	1	1.27	Hepatocellular
Estrogens	Tibolone	1	1.27	Hepatocellular
Antispasmodics	Alverine citrate	1	1.27	Hepatocellular
	Promethazine/dextromethorphan (supplement)	1	1.27	Hepatocellular