Portal de Periódicos da CAPES

What is idiosyncratic hepatotoxicity? What is it not?

2008; Lippincott Williams & Wilkins; Volume: 47; Issue: 6 Linguagem: Inglês

10.1002/hep.22332

1527-3350

John Senior,

Drug Transport and Resistance Mechanisms

In this issue, Lammert and colleagues1 challenge the "traditional" belief that idiosyncratic drug-induced liver injury (DILI) is not dose-related. They build upon the 1999 Uetrecht2 observation: "drugs given at a daily dose of 10 mg or less are rarely if ever associated with a high incidence of idiosyncratic drug reactions." Uetrecht postulated that idiosyncratic drug reactions did not occur at low doses because only limited amounts of reactive metabolite can be formed from them. DILI, drug-induced liver injury. Seeking data to investigate this concept further, Lammert and colleagues found among 233 most commonly prescribed oral medications that significantly higher incidences of severe DILI (liver failure, transplantation, or death) were reported for exposures to drugs with recommended daily doses of 50 mg or more. Repeat analyses of Swedish data3 have shown that 77% of 598 cases of DILI occurred in patients taking drugs at daily doses of at least 50 mg (higher doses), 14% at doses of 11-49 mg (intermediate doses), and only 9% for drugs at doses less than 10 mg (lower doses). Occurrences of liver transplants or deaths were 13% in patients taking higher doses, 9% at intermediate doses, and 2% at lower doses. Between 1990 and 2002, 137 liver transplantations were performed in the United States for DILI not caused by acetaminophen. After excluding the 26 patients with liver failure not due to oral prescription drugs, Russo and colleagues4 noted that 101 of the remaining 111 transplantations were performed in patients who took drugs at the higher doses, compared with 8 after intermediate doses and only 2 after lower doses. Based on these findings, Lammert and colleagues1 concluded that "some relationship may exist between daily dose of a medications and its propensity to exhibit hepatotoxicity" that is "worthy of further investigation." These results are notable, but the argument that idiosyncratic hepatotoxicity is independent of dose is something of a straw man in that what is being attacked is a misconception, though it may be widely believed. In their thoughtful reviews, both Uetrecht5 and Kaplowitz6 point out that it is wrong to say that idiosyncratic drug reactions are not dose-related. Uetrecht states: "Idiopathic drug reactions are often referred to as dose independent; this is incorrect and misleading." Kaplowitz wrote: "A common belief is that idiosyncratic allergic hepatotoxicity is unrelated to dose. This is a misconception." At a single-topic conference on DILI in 2005, none of the 24 invited expert speakers was reported to have stated that idiosyncratic DILI is not dose-related, although it was mentioned that idiosyncratic drug reactions were poorly understood, unpredictable, and probably the result of an interplay of genetic and nongenetic factors that make rare individuals susceptible to drug injury.7 Many authors have drawn inspiration from ideas developed by the late Hyman J. Zimmerman,8, 9 who made a distinction between true-predictable-intrinsic and idiosyncratic hepatotoxins. However, he also emphasized that this dichotomy is overly simplified, and the potential for producing injury forms a spectrum between the two poles. He very clearly defined idiosyncrasy as host-related, rather than an exclusive property of the drug or chemical to which a person might be exposed, or the dose thereof. Ulrich10 recently noted the complexity of risk factors, those of the drug and dose, and those of the individual patient. The key concept is that idiosyncratic liver injury, or injury to any organ, is not merely a function of the dose–duration of exposure to the drug or chemical but is also related to the unusual response of the person (or animal) that receives the agent and reacts to it. This concept is not new, but dates back 2400 years to the writings of Hippocrates.11 It is obvious to everyone that various people differ because genetic and other experiential differences between people may be considered natural phenomena.12 Finally, adequate description of idiosyncrasy so that others skilled in the art might be able to understand and use the concept is difficult and contentious. It may, however, be quite useful if research focuses on the differences between individual patients who react adversely to ordinary doses of drugs and the majority who do not, using the tools of metabolomics and proteomics, in addition to genomics. Tracking back the concept of differences between individual people to its origins in human thought takes us to the Hippocratean postulation of the four humors. These may perhaps relate to the four elements (air, fire, water, earth),13 of which everything was thought to be made in ancient times. The power of the concept of humors continued to Roman times, through Galen and many others, and persisted in medical vogue until the end of the eighteenth century, even up to the writings of Carl Jung in the early twentieth century and the Myers-Briggs Type Inventory personality types14 still in popular use today (Table 1). In the latter half of the twentieth century, an even more powerful concept of how people differ arose: that of helical DNA15 composed of a 3-billion-nucleotide sequence inherited at conception to constitute the human genome. Although approximately 99.9% of the sequences are common to all humans, even the 0.1% that are not allow for very great variation. If only variant nucleotides are considered, the theoretical 43,000,000 or 101,806,180 possible different sequences represent an extremely large number16 between a googol (10100) and a googolplex (1010∧100), which far exceeds the up to 125 billion humans estimated to have ever lived17 or who may live for eons to come. If the variant nucleotides are dispersed throughout the genome, even vastly greater numbers of sequences are possible. We know relatively little as yet about how the differences in nucleotide sequences may be expressed in people as differences in phenotype, or how environmental influences may modify sequence expressions. Interindividual genome sequence variations may also result from polynucleotide deletions or rearrangements, and DNA reading frame coding for peptide sequences may change. To return to the issue in question, what may all this have to do with idiosyncratic responses to drugs? Traditional toxicologic theory is heavily focused on dosing and effects in the average person or animal, and not on differences in response to the same dose by different individual animals or people. Simply to dismiss those who respond differently as "statistical outliers" and to seek only the average responses of a group of similar recipients may lose the game of understanding idiosyncrasy. A shift in perspective is needed to concentrate on what may be different about people (or animals) who are not identical and who respond differently than most subjects to the same stimuli. This paradigm was successful in past decades when genetically inherited abnormalities in phenotype were investigated to shed light on the enzymatic mechanisms of the abnormalities.18 Idiosyncrasy, derived from the Greek ιδιOς (idios = one's own) + συγ (sug or syn = together) + κρασις (crasis = mixture) was the ancient Greek word for the Hippocratean idea of individual mixtures of four humors that determined a person's type, a word still in use by modern Greeks meaning a person's temperament or constitution. It applies not only to adverse reactions (such as DILI) but also to beneficial therapeutic effects, and to responses to stimuli other than chemical ones, as seen in people's temperamental patterns of responses to situations. It does not necessarily mean "rare" but implies only that it may not be seen in most people; it is not specified what the numeric value of "most" is, whether 2 of 3 or 999,999 of 1,000,000. Susceptibility to high doses of alcohol is idiosyncratic, occurring in perhaps only 15%–20% of heavy drinkers who develop alcoholic hepatitis and are likely to progress to chronic alcoholic liver disease and cirrhosis upon repeated exposures over the years. "While excessive alcohol consumption is clearly associated with the development of liver disease, advanced liver disease occurs in only a minority of heavy drinkers."19 Isoniazid causes alanine aminotransferase elevations in about 20% of people taking it for the first time, but "most" (all but 0.1%–0.2%) can adapt and later tolerate the usual therapeutic doses for prevention of tuberculosis.20 Idiosyncratic does not mean unpredictable or unexplained, although it usually is because we do not understand it well enough to make predictions about in whom it will occur. It also does not mean unrelated to dose, because even those who are idiosyncratically susceptible may show a dose relationship, albeit at lower dose ranges than most. Idiosyncrasy refers to individual differences in responses to stimuli, which we need to study and understand better using all the tools of systems biology that have recently been developed, as mentioned above. The study of outliers, exceptions to the rules, may lead to better understanding of mechanisms by which DILI is caused. This is not an idle exercise in semantics but a distinction of potentially great importance in unraveling the mysteries and challenges of understanding the mechanisms by which DILI may be caused, and for possible predictive identification of individuals who should not be exposed to certain drugs. Lammert and colleagues have done a service in reminding us that we need to be careful how we talk, write, and think about concepts such as idiosyncrasy that may seem simple but are deceptively elusive and complex.