MONITORING LIVER DAMAGE

February 18, 2002

Experts give clues on how pharmacists can prevent drug-induced liver toxicity

 

COVER STORY

MONITORING LIVER DAMAGE

Experts give clues on how pharmacists can prevent drug-induced liver toxicity

Drug-induced liver disease is a significant problem. It is the most common cause of acute liver failure, and it accounts for one-third to one-half of all reported cases. Literally hundreds of drugs can cause hepatotoxicity—though it is usually a rare side effect.

Drug-induced hepatotoxicity made headlines in 2000 when Rezulin (troglitazone, Parke-Davis) was withdrawn from the market, demonstrating that even a side effect once considered somewhat rare could become a major problem.

Among the more than 1.5 million people with diabetes who used troglitazone, at least 63 deaths tied to liver problems have been blamed on the drug. In addition, more than 4,000 troglitazone lawsuits are expected to go before juries. Based on these figures, roughly 0.26% of patients may have experienced serious adverse effects—but this was a large enough, and serious enough, problem to cause troglitazone to be withdrawn from the market.

The impact of troglitazone is evident on many levels. Neil Kaplowitz, M.D., director and principal investigator of the University of Southern California Research Center of Liver Diseases (supported by the National Institute of Diabetes, Digestive, and Kidney Diseases) and professor, department of medicine, University of Southern California, noted, "I think that the Food & Drug Administration is very sensitive, as are the various advisory groups, in looking at liver changes and recognizing that what seem like minor changes may be an early signal that something needs to be paid attention to." He also cited the FDA's employment of liver experts and the sensitization of pharmaceutical companies to drugs that could cause toxicity. "Most of the companies now have advisory boards of liver experts, so whenever they see a liver problem, they bring in an advisory board of outside expertise to give them advice as to the significance of what they're seeing," he said.

Since troglitazone has been pulled, liver toxicity has emerged as an issue for other drugs/supplements as well. Consider the following products:

• Nefazodone (Serzone, Bristol-Myers Squibb) recently added a black box warning on its package insert, stating that cases of life-threatening hepatic failure have been reported and patients should be alerted to signs of liver dysfunction (jaundice, anorexia, gastrointestinal complaints, malaise) and should report them to their physician immediately if they occur.

• Hepatic reactions have caused the European Medicines Evaluation Agency and Health Canada to recommend revision of leflunomide (Arava, Aventis) labeling to reflect 296 cases of hepatic reactions (129 serious, nine ending in death).

• Because of reported liver problems, the Centers for Disease Control & Prevention and the American Thoracic Society issued revised guidelines for latent tuberculosis, limiting the circumstances in which the combination of rifampin and pyrazinamide is prescribed, as well as strengthening the level of patient counseling and monitoring recommended.

• A comprehensive retrospective review of more than 21 different AIDS clinical trials confirmed that antiretroviral therapy is associated with a high rate of severe hepatotoxicity, regardless of drug class or combination.

• The FDA sent a letter to healthcare professionals requesting assistance in determining whether dietary supplements containing kava are associated with liver toxicity.

In November 2002, the FDA, Pharmaceutical Research & Manufacturers of America, and American Association for the Study of Liver Diseases issued a white paper on the current status of postmarketing surveillance for drug-induced liver toxicity. They also set directions and recommendations for discussion at a public workshop last February. Slides of presentations made at this workshop are available on-line at www.fda.gov/cder/livertox/.

How liver toxicity presents itself

The predominant presentation of liver toxicity is acute hepatitis or cholestatic liver disease, but it can mimic all forms of acute and chronic hepatobiliary diseases. An acute hepatitis reaction is characterized by symptoms of jaundice; markedly elevated serum transaminase levels; and, in severe cases, coagulation disorders and encephalopathy that indicate acute liver failure. It may resolve over several weeks. Cholestatic disease is not usually life-threatening, but it has symptoms of jaundice, disproportionately increased serum alkaline phosphatase levels, and itching. It may resolve very slowly over months.

Liver toxicity is not well understood, and there are many factors involved in its pathogenesis. There may be a characteristic pattern either for a drug, class of drugs, or other factors that predicts an individual's susceptibility. Or, because these reactions occur in a very small percentage of patients taking the drugs, perhaps genetic, environmental, or a combination of factors predispose certain individuals to liver damage. Differences in susceptibility may result from the interplay of genetic polymorphisms of various systems of toxicity, defense, drug metabolism, and repair. Alternatively, individuals might be predisposed to liver toxicity because of patterns of gene expression.

Some drug-induced hepatotoxicity is predictable (high incidence), but other hepatotoxicity is unpredictable (low incidence). Most adverse events are unpredictable. Unpredictable adverse events can be categorized as hypersensitivity or idiosyncratic reactions. With hypersensitivity reactions, "there seems to be an immune response by the liver that's drug dependent. It's probably related to a metabolite being the immunogen," stated Kaplowitz. This sort of reaction occurs early, usually during the first month or two of taking the drug, and it includes features of allergy.

Idiosyncratic reactions have a more "cryptic pathogenesis." These reactions occur in a small percentage of individuals, and the action can be delayed for many months. "Patients can be doing perfectly fine with normal liver tests, and then seven or eight months later, they suddenly develop acute liver disease," Kaplowitz said. He considers it "the holy grail" of the field to understand the idiosyncratic reaction.

What to do when it occurs

Liver toxicity is deterred whenever possible by discontinuation of the drug. A common (but not uniform) recommendation for when to stop drug therapy is when liver enzymes climb to more than three times the upper limit of normal. Kaplowitz stated that there are no accepted "antidotes" for liver toxicity, aside from acetaminophen poisoning, where N-acetylcysteine is used, because it generates or promotes the production of glutathione that detoxifies the acetaminophen toxic metabolite.

"In the subgroup of individuals who have a very severe hypersensitivity reaction with features of allergy with severe rash and so on, as might occur with phenytoin, there might be some merit in using a short course of steroid therapy," Kaplowitz noted. "But that's never been agreed upon or studied."

Drug-induced hepatotoxicity is not a new problem. John Senior, M.D., of the Office of Post-marketing Drug Risk Assessment at the FDA, labels hepatotoxicity the most common single adverse effect causing major drug problems—i.e., withdrawal after marketing or nonapproval. Examples supporting his contention date from 1956 with iproniazid (Marsilid) and continue with ibufenac, ticrynafen (Selacryn), benoxaprofen (Oraflex), perihexilene, dilevalol, bromfenac (Duract), and troglitazone. In addition, a partial list of drugs that are relegated to second-line status or bear serious warnings include pemoline, tolcapone, trovafloxacin, felbamate, valproic acid, ketoconazole, nicotinic acid, acetaminophen, chlorzoxazone, isoniazid, dantrolene, rifampin, zileuton, tacrine, labetalol, and diclofenac.

According to Senior, there are two pressing problems of drug-induced hepatotoxicity. The first problem, for the FDA, is how to detect hepatotoxic effects and the need to withdraw approved drugs. The second problem occurs in medical practice: The incidence of acute liver failure is too high, how can it be prevented? Retrospective analyses of clinical trials of drugs later found to cause liver toxicity suggest there are signals that warn of toxicity. But, because most trials are short-term (less than one year) and are tested in a population too small to show overt toxicity (which is extremely rare), signals may be the only indication of potential toxicity.

Kaplowitz explained, "Signals tend to occur in layers. The first [and probably most sensitive signal] is the increased incidence in serum ALT abnormalities." Typically in some adult study populations, between 0.1% and 0.6% of the population will have transient serum ALT increases to more than three times the upper limit of normal. "When you start seeing an incidence that's two to three times that, it suggests that [a drug] might have a propensity to cause injury. That doesn't always mean there is going to be serious injury, but it is a first signal." He added that some drugs, like the statins, tend to have that signal, but they very rarely cause overt liver disease.

The next layer is more severe ALT abnormalities. "When you start getting up to eight to 10 times the upper limit of normal in a study population, that's a more significant signal to me, because that doesn't happen very often, just randomly—almost never," Kaplowitz observed. To the best of his knowledge, no one in the troglitazone control group had an ALT eight to 10 times the upper limit of normal—but some of the patients studied on active drug did. "When you start getting into that range, you're looking at something that shows a propensity for more serious damage," he said.

"An unequivocal signal is the appearance of those abnormalities—higher transaminases accompanied by hyperbilirubinemia or jaundice—that mean a much more severe liver insult," Kaplowitz remarked. When that occurs, the liver injury is significantly severe and there's a potential for risk of developing life-threatening liver disease.

Kaplowitz continued: "Then if you invoke the so-called Zimmerman's rule—which is simply his observation that with a variety of drugs that cause the toxic effect on the liver, there's always this relationship that about 10% of people who develop overt liver disease with jaundice and high transaminases either die or need a liver transplant"—you can predict the risk of death without liver transplant. "When you apply all of that to the Rezulin case, as an example, the numbers all work out very well," he said.

As such analyses have evolved, the FDA has incorporated liver toxicity assessment into drug approvals and withdrawals. But the issue is far from clear-cut. Julie Beitz, director, division of drug risk evaluation, Center for Drug Evaluation & Research at the FDA, commented, "The issue of drug-related liver toxicity has to be evaluated on a case-by-case basis. Many drugs can cause elevations of liver function tests, but rarely is severe liver injury identified."

Robert Temple, M.D., Office of Medical Policy Director for the FDA, explained how the FDA uses Zimmerman's rule to predict severe liver injury. "Translated to current practice, we look for an increased rate of three times, five times, eight times, etc., elevated transaminases with even very few cases of elevated bilirubin," he said. "That finding, we think, predicts major trouble. Note, if you had a rate of 1:1,000 cases that might predict a rate of 1:10,000 serious liver injuries. A drug with such a finding might not be approved at all [e.g., dilevalol some years ago], or it might be approved with warning language. Once a drug is marketed, you don't need the signal; you either do or don't see evidence of hepatotoxicity."

When evidence of liver toxicity becomes evident, it is necessary to ascertain whether the drug is the causative agent. According to Beitz, causality in spontaneous adverse event reports is assessed by the temporal relationship between the event and the administration of the drug, whether the event subsided when the drug was stopped, whether the event returned when the drug was resumed, and other factors.

If a drug is found to cause liver injury, the FDA then must evaluate whether to allow it to continue to be marketed or to pull it from the market. Beitz stated, "The decision is very complex and involves consideration of many factors, including the seriousness of the event, the strength of the association between the adverse event and the drug in question, whether or not prescribing or dispensing behaviors can be altered in some way to improve safety, and whether there are alternative drugs that patients could use in the event the drug in question is withdrawn."

Does monitoring work?

Some drugs that can cause liver toxicity continue to be marketed, albeit with black box warnings, limited accessibility, or monitoring requirements. While monitoring is very important, Beitz commented that the FDA is aware of how difficult it is for monitoring requirements to be carried out even when they are prominently labeled.

Kaplowitz said that compliance with monitoring is poor. He also pointed out the more ominous problem, "The monitoring approach works pretty well in circumstances where the liver disease evolves late and slowly. But it doesn't always. In the case of Rezulin, one of the problems was that some of the patients—I don't know how many—had monthly monitoring and, at a certain month, had a normal liver test, and two weeks later were in the hospital with acute liver failure. So monthly monitoring did not prevent that from happening."

"It's never been proven that monitoring actually works, even if it's adhered to," Kaplowitz added. "The presumption [of monitoring] is that you're going to identify the same process that sometimes will progress and sometimes won't. But it's alternatively possible that the severe reactions have a mechanistic basis that is completely different from these mild reactions. It's just not known or proved." He does believe monitoring is the best option we currently have and that recommended monitoring schedules should be followed. "I think that, empirically, the serum ALT is a sensitive test that is a reasonable signal," he said.

So, how should pharmacists deal with the issue of drug-induced liver toxicity? William Kirchain, Pharm.D., BCPS, Medical University of South Carolina, remarked, "Those pharmacists involved in direct monitoring can add another level of care by being diligent in ordering or recommending baseline and follow-up liver enzyme labs on patients taking drugs with predictable liver toxicity."

Bruce Carlstedt, Ph.D., professor of pharmacy practice, Purdue University, noted that it is important for pharmacists to be involved by having "awareness of drugs likely to cause [liver toxicity], detecting it in patients for whom it has not been considered, and—perhaps most important—reporting it in ongoing adverse drug reaction surveillance or to the FDA."

Kaplowitz believes that when it's unequivocal that a drug has been associated with liver injury, pharmacists need to warn the patient—with the appropriate qualification, particularly when liver toxicity is rare and not expected. He considers it important that patients know that "if some change in symptoms occurs—e.g., they start feeling weak or fatigued or have nausea, vomiting, abdominal pain, any systemic type of symptoms, let alone jaundice or dark urine—they should immediately notify their physician to address the question of whether they're having an adverse effect."

Kirchain commented, "The most important role for all pharmacists is to specifically ask patients when they are obtaining refills if they are experiencing any problems. The key to survival in hepatic toxicity is early detection."

Early identification of symptoms "is a very useful finding—with some drugs more than others," Kaplowitz added. "A good example of that is isoniazid, where public health agencies tend to prescribe the drug in one-month allotments so the patients come in, get a prescription, come in a month later. They are informed about possible liver adverse effects, and they're asked whether they have had any of those symptoms. That approach alone, even without doing serum ALTs seems to be very effective in avoiding serious outcomes," he said.

Much more research is needed to better understand drug-induced liver toxicity and to prevent it. In the aftermath of troglitazone, attention is being focused on this serious side effect and approaches to deal with it. Pharmacists can play an important role in safeguarding patients' health by suggesting and monitoring liver tests (when their practice setting permits); educating and reminding patients about the occurrence of symptoms that might indicate toxicity; watching for indications that liver toxicity is present; and, if it is, making sure that it is reported, so that postmarketing surveillance accurately tracks types of adverse events and frequency of occurrence.

Kathy Hitchens, Pharm.D.

The author is a clinical writer in the Indianapolis area.

SELECTED DRUGS THAT CAN CAUSE LIVER TOXICITY

acarbose

acetaminophen

allopurinol

amiodarone

amoxicillin/clavulanic acid

atorvastatin

azathioprine

carbamazepine

chlorzoxazone

cyclophosphamide

cyclosporine

dantrolene*

diclofenac*

disulfiram*

efavirenz

erythromycin

felbamate*

flutamide*

gemfibrozil

isoniazid*

itraconazole

ketoconazole*

labetalol*

lovastatin

methotrexate

methyldopa

minocycline

nefazodone*

nevirapine

niacin/nicotinic acid*

nitrofurantoin

pioglitazone

pravastatin

pemoline*

phenytoin

protease inhibitors

pyrazinamide

ranitidine

rifampin

ritonavir

rosiglitazone

simvastatin

stavudine

sulfonamides

sulindac

tacrine

tamoxifen

terbinafine*

tetracycline

tolcapone*

trovafloxacin*

trazodone

valproic acid*

vitamin A

zafirlukast*

zileuton

 

SELECTED HERBS REPORTED TO CAUSE LIVER TOXICITY

Chaparral

Comfrey

Germander

Kava*

Pennyroyal

Red yeast

Valerian

*Subject of FDA letter to healthcare professionals 12/19/01 (www.cfsan.fda.gov/~dms/ds-ltr27.html ).

 

Kathy Hitchens. MONITORING LIVER DAMAGE. Drug Topics 2002;4:46.