Methods

Top Methods Results Discussion Author & Article Info References

The FDA's Adverse Drug Event Reporting System ("Spontaneous Reporting System") contains approximately 600 000 reports of adverse drug events submitted since 1969. Reports of adverse events are submitted principally by physicians and pharmacists to pharmaceutical companies who must, by law, forward the information to the FDA.

A computer-assisted search of reports received by the FDA from February 1989 through March 1991 in which flutamide was listed as the suspect drug was done using 33 terms indicative of hepatotoxicity. We excluded foreign, literature, and consumer reports and only reviewed reports that originated in domestic patient care settings. We also excluded duplicates and paired initial and follow-up reports. This search yielded 49 reports that included increased liver enzyme levels, hyperbilirubinemia, jaundice, nausea, vomiting, and cholestatic hepatitis. We reviewed reports that met the following case definition for hepatocellular reactions possibly associated with flutamide: an aspartate aminotransferase (AST) or alanine aminotransferase (ALT) level of more than six times the upper limit of normal; a compatible time sequence; and the lack of a more plausible infectious, obstructive, toxic, or other drug-induced explanation. Of the 49 reports, 21 did not meet the case definition due to decreased liver enzyme levels or failure to mention them (most of these patients with follow-up information had resolution of liver enzyme abnormalities after drug discontinuation). One patient's problems resolved without drug discontinuation, and four patients had liver metastases, cholelithiasis, or infections that may have been responsible for increased liver enzyme levels. We included one report that was subsequently published [11].

Questionnaires were sent to reporting physicians or pharmacists requesting additional information on the 22 patients, such as alcohol use, exposure to anesthesia and blood transfusions, diethylstilbestrol and analgesic use, as well as results of studies to exclude diagnoses of viral hepatitis, liver metastases, and biliary obstruction. One reporter did not return his questionnaire, 2 reporters indicated no further information was available, and 19 returned the questionnaires with additional information. One patient who continued receiving flutamide with resolution of abnormal liver function test results was not included.

Results

Top Methods Results Discussion Author & Article Info References

The characteristics of the 19 patients are summarized in Table 1. The following two case reports show the evolution of flutamide hepatotoxicity, its detection, and its response or lack of response to clinical intervention.

Patient 1

A 62-year-old white man was found to have prostatic carcinoma in November 1990 and was prescribed flutamide and an LHRH analog. He was hospitalized on 28 January 1991 with a history of coronary artery disease, hypertension, and gradually worsening jaundice of 3 weeks duration, in addition to the prostate cancer. Flutamide was discontinued on admission. The patient had no history of transfusion, intravenous drug use, recent travel, or toxic exposure. His other medications included diltiazem hydrochloride and nitroglycerin. The patient had normal results of right upper-quadrant ultrasound, increased AST and ALT levels, and negative test results for hepatitis surface antigen, antibodies to hepatitis B surface and core antigens, and hepatitis C virus. Test results for antibody to hepatitis A virus were positive, but results for hepatitis A virus immunoglobulin M (IgM) were negative. Physical examination indicated that the patient was neurologically intact; however, a mental status examination indicated decreased concentration.

On hospital day 2, his mental status deteriorated, he was transferred to the intensive care unit for worsening encephalopathy, which progressed to stage IV; he became deeply comatose with grand-mal seizures. His condition deteriorated; his aminotransferase enzyme levels peaked on hospital day 2 (AST, 49.6 µkatal/L; ALT, 31.2 µkat/L, and -glutamyltransferase, 5.13 µkatal/L). On day 9, his bilirubin level peaked at 605 µmol/L (35.4 mg/dL), and then his prothrombin and partial thromboplastin times increased to 45 seconds and 100 seconds, respectively. The patient had multiple-organ failure, including renal failure and pneumonia, and on day 12 the patient died. Autopsy findings indicated massive hepatic necrosis (weight, 760 g), and microscopic examination indicated widespread necrosis, bile duct proliferation, and focal macrovesicular steatosis (Figure 1).

View larger version (153K): [in this window] [in a new window]   Figure 1. Histologic findings in the liver. Massive liver necrosis with bile duct proliferation and focal macrovesicular steatosis. (Hematoxylin and eosin stain; original magnification, x 40).

Patient 2

A 79-year-old man who was previously diagnosed with localized adenocarcinoma of the prostate was referred on 18 July 1990 to a gastroenterologist for evaluation of jaundice and increased liver enzyme levels. Three months earlier, in April 1990, he started treatment with an LHRH analog, leuprolide acetate (Lupron, TAP Pharmaceuticals Inc., Deerfield, Illinois) and flutamide, 750 mg daily. Leuprolide was given as a 1-mg subcutaneous injection administered in April, May, and June. During the 2 months before presentation, he noted progressive weakness and fatigue. No history existed of jaundice, liver disease, alcohol abuse, tattoos, intravenous drug abuse, toxic chemical exposure, or easy bruising, bleeding, or itching.

In February 1990, liver function test results were normal. On 9 July, the ALT level was 16.7 µkatal/L, the AST level was 15.4 µkatal/L, and the total bilirubin level was 27 µmol/L (1.6 mg/dL). By 16 July, the ALT level was 18.3 µkatal/L, the AST level was 17.7 µkatal/L, and the bilirubin level was 36 µmol/L (2.1 mg/dL). Alkaline phosphatase levels were normal. Leuprolide and flutamide were discontinued on 18 July. A computed tomography scan showed an upper normal-sized liver, degenerative joint disease of the lumbar spine, and a cyst on the right kidney. An ultrasound scan showed a normal biliary tree. His physical examination was unremarkable with no evidence of spider angiomata, palmar erythema, asterixis, organomegaly, or ascites. Hepatitis serologic test results were negative for hepatitis A and B, and his prothrombin time was normal. An antinuclear antibody titer (>1:640) was positive.

Aminotransferase enzyme levels peaked on 26 July, with an AST level of 32.0 µkatal/L and an ALT level of 30.7 µkatal/L; bilirubin levels peaked on July 30 at 92 µmol/L (5.37 mg/dL). Gradually, the patient's symptoms improved and liver chemistry test results returned to normal by 16 October. The antinuclear antibody titer was repeated in December 1990; it was 1:320, suggesting that the positive titer was a coincidental finding. The patient was feeling well and subsequently had an orchiectomy.

Survey Results

Of the 19 patients for whom we had additional information, two patients received flutamide for benign prostatic hypertrophy (for Investigation of a New Drug or off-label use) and 17 patients were receiving flutamide for prostate cancer. All patients received the standard dose of 750 mg per day except for one patient with benign prostatic hypertrophy who received 1500 mg per day. Excluding the two 58-year-old patients treated for benign prostatic hypertrophy, the median age of patients was 71 years. The median duration of flutamide treatment for all 19 patients was 56 days. By definition, all patients had abnormal liver function test results. Clinical jaundice was reported in 11 patients, hyperbilirubinemia in 16, and prolonged prothrombin time in 4. Encephalopathy was reported in 3 patients and renal failure in 4. Ten patients who did not require hospitalization recovered after discontinuing flutamide or decreasing the dose. One of these patients was rechallenged with 750 mg of flutamide after a 1500-mg dose had been discontinued 2 months earlier due to increased liver enzyme levels. Five days later, the AST level was 1.5 µkatal/L, the ALT level was 3.1 µkatal/L, and the alkaline phosphatase level was 4.7 µkatal/L. The patient again recovered after drug discontinuation.

Four patients were hospitalized and, after flutamide was discontinued, they recovered. Five patients, all of whom had slight decreases or no improvement in liver enzyme levels after flutamide discontinuation, died during hospitalization; liver biopsies were not done in any of these patients. Of the five patients who died, three were autopsied. Postmortem examination of the liver of one patient [11] showed massive hepatocellular necrosis with collapse of lobules, ductal proliferation, and severe cholestasis; no inflammatory cells were seen. The findings were interpreted as possibly drug-related. No residual carcinoma of the prostate was seen in the vertebrae of this patient who had multiple bone metastases before orchiectomy and flutamide therapy. Liver autopsy results of the second patient are described in the first case report. The third patient also had evidence of massive hepatic necrosis (weight, 700 g) at autopsy with extensive hemorrhagic necrosis, moderate focal bile stasis, and bile-duct proliferation.

Our survey of reporters showed that none of the patients were known to be heavy drinkers of alcoholic beverages; none had exposure to general anesthesia within 1 month of abnormal liver function test results; no blood transfusions, diethylstilbestrol exposure, or known use of prescribed or over-the-counter analgesics were reported. Serologic test results for viral hepatitis, including four studies for hepatitis C antibody, were done in 14 patients. All tests were negative except in four patients who were reactive to hepatitis A antibody. Further testing of one patient showed that IgM antibody was negative. Computed abdominal tomography and ultrasonography scans were negative for biliary obstruction and liver metastases. Five of 19 patients received no other medications besides flutamide. Six patients received an LHRH analog only. One patient had chronic theophylline use, another had chronic allopurinol use, another used gemfibrozil, and a fourth had chronic diuretic use. Three patients, all of whom died, received another drug in addition to flutamide and an LHRH analog. These medications included metoclopramide (whose labeling states that it infrequently causes hepatotoxicity when administered concurrently with hepatotoxic drugs [12]) and methadone in one patient; diltiazem hydrochloride and nitroglycerin in a second patient; and inhalant bronchodilators in the third patient. Viral hepatitis studies were negative for the patient using methadone.

In addition to the five deaths that occurred in U.S. patients, the manufacturer reported to the FDA death possibly due to flutamide in an Austrian patient. This patient, a 56-year-old man treated with an LHRH analog and flutamide since June 1989, was hospitalized in November 1989 for a radical prostatectomy. On admission, the patient was clinically jaundiced; the AST level was 6.7 µkatal/L, the ALT level was 3.0 µkatal/L, and the bilirubin level initially was 325 µmol/L (19 mg/dL), and it increased to 855 µmol/L (50 mg/dL). The patient subsequently developed hepatic and renal failure. Flutamide was continued until the patient died on 30 November 1989. The autopsy showed a small liver (15-cm diameter) with some granulomas and necrotic areas. No microscopic metastases were evident.

Discussion

Top Methods Results Discussion Author & Article Info References

We summarized reports indicating that the antiandrogen drug, flutamide, possibly causes hepatotoxicity. Between its marketing in early 1989 through March 1991, the FDA received 49 reports of flutamide possibly causing abnormal liver enzyme levels, hyperbilirubinemia, nausea, vomiting, jaundice, and cholestatic hepatitis. Of these, 22 reports met our case definition for severity of the reactions, appropriate temporal sequence, and lack of other known causes. Of the 19 patients in whom additional information was obtained, 14 patients with serious liver toxicity had resolution of abnormalities after discontinuation or dose lowering of flutamide, but five patients died from progressive liver disease. In three patients, autopsies showed massive hepatic necrosis [11] similar to that for the Austrian patient described above, whose flutamide dose was not discontinued. In most cases, thorough workups to exclude other possible causes (for instance, metastases, obstruction, infection, excess alcohol, and other drugs) were negative.

Reports of Hepatitis

Reversible hepatitis in flutamide users has been reported [7-10, 13-15]. In 1987, Lundgren [13] reported that toxic liver damage occurred in 1 of 10 patients treated with flutamide. In 1988, Danish investigators [14] reported the reversal of severe hepatotoxicity after discontinuation in 1 of 20 patients treated with flutamide. In 1989, Hart and Stricker [7] in the Netherlands reported the reversal of a case of cholestatic hepatitis after discontinuation of flutamide therapy. Robinson and colleagues [15] found that 5 of 149 (3.4%) patients treated with an LHRH analog and flutamide had liver toxicity compared with no cases of 148 patients treated with orchiectomy (Fisher exact test, two-tailed, P = 0.06).

Four additional cases of hepatitis have been reported from France [9, 10] and Denmark [8]. A 60-year-old Danish patient had severe hepatic insufficiency with jaundice, ascites, and hepatic coma after treatment for 16 weeks with flutamide, 750 mg per day. Microscopic analysis of a fine-needle liver biopsy specimen showed nonspecific parenchymal changes with several focal necroses and changes consistent with long-standing cholestasis (bile thrombi in rosettes and bile pigment in Kupffer cells). Portal tract changes indicating extrahepatic obstruction were absent. After flutamide discontinuation, the clinical condition of the patient improved, liver chemistry test results normalized, and on follow-up 2 years later, the patient had no symptoms of liver disease. A 69-year-old Danish patient with poorly differentiated adenocarcinoma developed jaundice, discolored stools and urine, itching, and diarrhea 12 days after initiation of flutamide therapy [8]. Flutamide was discontinued; liver chemistry test results improved and 2 months later the jaundice disappeared.

Flutamide-associated Hepatotoxicity

The temporal association between exposure and the onset of abnormal liver function test results as well as the exclusion, in most cases, of other causes indicates that flutamide can cause severe hepatotoxicity. An LHRH analog recommended to be administered with flutamide [12] does not appear to cause the hepatotoxicity [8]. In our study and several others [13, 14, 16], hepatotoxicity developed in patients treated with flutamide and without the analog. In addition, when flutamide was promptly discontinued even when the analog was continued, abnormal liver function test results normalized and clinical symptoms resolved in most patients. In at least one patient in whom flutamide was not discontinued, liver disease progressed and resulted in the patient's death. Also, rechallenge with flutamide in one patient was associated with increased liver enzyme levels. Finally, liver biopsy and autopsy findings have also been consistent with drug-induced liver damage.

The incidence of hepatotoxicity cannot be estimated from spontaneous reports due to the unknown magnitude of under-reporting of adverse drug reactions [17]. Also, the number of patients treated with flutamide is unknown, but the volume of outpatient prescriptions was estimated to be 99 000 in 1989, 248 000 in 1990, and 313 000 in 1991 [18]. The manufacturer's labeling for flutamide states that hepatitis and jaundice occurred in fewer than 1% of patients treated with flutamide and an LHRH agonist in a multicenter clinical trial [12]. Because several studies have been published as abstracts, with little information concerning study design, or had incomplete ascertainment of liver enzyme levels in patients, estimates of incidence are difficult but they range from less than 1% to about 5% [3, 4, 14-16].

The biological mechanism of flutamide-induced hepatotoxicity is unknown. Hart and Stricker [7] speculated that the cholestatic hepatitis associated with flutamide in their patient might have been caused by testosterone-mediated cholestasis, but repeated eosinophilia levels of 10% to 16% were also compatible with an immunoallergic type of reaction. Moller and colleagues [8] suggested that the hepatic injury in their two patients was idiosyncratic—possibly due to interference with metabolic processes in the hepatocytes, destruction of cells through a toxic effect on essential structures, or induction of an immunologic reaction leading to necrosis and cholestasis. No fever or rash were mentioned, and eosinophilia was the only sign of hypersensitivity that occurred [19]. Our patients appeared to have massive hepatocellular necrosis rather than an allergic mechanism.

Conclusion

Whatever the mechanism, the revised manufacturer's labeling [12] states that "since transaminase abnormalities, cholestatic jaundice, hepatic necrosis, and hepatic encephalopathy have been reported with the use of flutamide, periodic liver function tests should be considered". It also warns that "there have been reports of death following severe hepatic injury associated with the use of flutamide". Physicians should not prescribe flutamide for benign prostatic hypertrophy until it has been approved by the FDA for that indication based on its safety and efficacy. Patients should immediately report symptoms such as anorexia, nausea, vomiting, fatigue, discolored urine, and jaundice and should discontinue flutamide.

Additional cases of flutamide-associated hepatotoxicity should be reported to the FDA's Office of Epidemiology and Biostatistics. To facilitate reporting, adverse drug event forms are available on the last page of recent versions of the Physician's Desk Reference, the FDA Medical Bulletin, the USP-Drug Information for the Health Care Professional, and selected editions of the American Medical Association's Drug Evaluations.

This paper contains the views of the authors and not necessarily those of the FDA.