- Email: [email protected]
Hepatotoxicity due to troglitazone: report of two cases and review of adverse events reported to the United States Food and Drug Administration
THE AMERICAN JOURNAL OF GASTROENTEROLOGY © 2000 by Am. Coll. of Gastroenterology Published by Elsevier Science Inc.
Vol. 95, No. 1, 2000 ISSN 0002-9270/00/$20.00 PII S0002-9270(99)00766-2
CASE REPORTS
Hepatotoxicity Due to Troglitazone: Report of Two Cases and Review of Adverse Events Reported to the United States Food and Drug Administration James Kohlroser, D.O., Jijy Mathai, M.D., James Reichheld, M.D., Barbara F. Banner, M.D., and Herbert L. Bonkovsky, M.D. Departments of Medicine, Pathology, Biochemistry and Molecular Biology, and The Liver, Biliary, Pancreatic Center of UMASS Memorial Health Care and the University of Massachusetts Medical School, Worcester, Machusetts
ABSTRACT Two patients (a 48-year-old woman and a 62-year-old man) developed clinical and laboratory signs of hepatotoxicity due to troglitazone (Rezulin), a thiazolidinedione used in treatment of diabetes mellitus. There was no clear clinical evidence of drug allergy, although the woman experienced colitis before the onset of recognized hepatotoxicity. Liver biopsies showed bridging necrosis and fibrosis in the woman and hepatitis with granuloma formation in the man. The abnormalities in liver chemistries resolved promptly upon cessation of the drug. Cases involving 46 patients reported to the United States Food and Drug Administration are also reviewed. Troglitazone is a useful new oral antihyperglycemic agent, but in about 1.9% of patients hepatotoxicity has occurred, which may be severe and even fatal. Frequent monitoring of serum liver chemistries in patients taking the drug is essential. (Am J Gastroenterol 2000;95: 272–276. © 2000 by Am. Coll. of Gastroenterology)
INTRODUCTION Troglitazone is a thiazolidinedione with insulin-sensitizing activities when administered to humans or animals with type 2 (noninsulin-dependent) diabetes mellitus. It has been shown to have a number of desirable metabolic effects on glucose and lipid metabolism, and it has been approved for clinical use in Europe, Japan, and the United States (1, 2). A major adverse effect of troglitazone is the development of hepatotoxicity. In early clinical trials, elevations of serum aminotransferases (⬎3 times the upper limit of normal) occurred in 48 of 2510 (1.9%) subjects receiving troglitazone (3). In December 1997 and again in August 1998, the United States Food and Drug Administration (FDA) issued stronger warnings and recommendations after receiving reports of ⬎100 cases of liver damage, including liver failure requiring transplantation in three patients and death in another patient treated with troglitazone (4). These events must be balanced against the beneficial effects of the drug on glucose and lipid metabolism and interpreted in the light
of the large numbers of patients who have been, and are being treated with it (⬎800,000 in the US and Japan). Concerns about the hepatotoxicity of troglitazone led the Medicines Control Agency of the United Kingdom to request voluntary withdrawal of the drug from the UK, effective December 1, 1997 (5). This request was honored by the pharmaceutical companies involved. In this article, we describe two previously unreported, carefully studied patients with troglitazone hepatotoxicity, comparing their features with other reported patients, including review of 46 cases of hepatotoxicity reported to the FDA, information on which we obtained under the Freedom of Information Act. Three case reports have recently been published reporting troglitazone induced hepatotoxicity. All three were reported to the FDA; two of the patients improved with discontinuation of troglitazone and supportive care, whereas the third patient developed hepatic encephalopathy and worsening liver failure requiring liver transplantation (5, 6) Because all identifying information is removed from the Medwatch reports before being released under the Freedom of Information Act, we cannot ascertain whether these cases are included in the reports that we obtained from the FDA.
CASE REPORTS Patient 1 A 48-year-old Hispanic woman developed jaundice 8 months after starting troglitazone for the treatment of diabetes mellitus type 2. Her initial dose was 200 mg b.i.d., which was increased to 400 mg b.i.d. after the first 3 months. Notably, her serum ALT had been normal 3 yr before beginning troglitazone (34 U/L), but was mildly elevated (44 U/L) after 1 month of therapy (Fig. 1A). She had no known drug allergies. She presented with diarrhea 5 months after beginning troglitazone and was found to have mild colitis by colonoscopy. Stool cultures for enteric pathogens, Yersinia, and Campylobacter were negative, as were examinations for ova
AJG – January, 2000
Figure 1. Summary of liver chemistries of two patients with troglitazone hepatotoxicity. A: Patient 1. B: Patient 2.
and parasites. Her diarrhea resolved without further intervention. She denied the use of other prescription or overthe-counter medications and rarely used alcohol. On examination, she was without lymphadenopathy; there were no stigmata of liver disease and there was no abdominal tenderness, mass, or hepatosplenomegaly. Liver chemistries revealed elevated serum aminotransferases, ALT 653 U/L, AST 413 U/L (normal ⱕ40 U/L), and total bilirubin 1.5 mg/dl (normal ⱕ1.2 mg/dl) (Fig. 1A). The complete blood count was normal with white blood cell count of 6700/l and a manual differential of 50% polys (polymorphonuclear leukocytes), 35% lymphocytes, 13% monocytes, 1.8% eosinophils, and 0.8% basophils. Serologies for hepatitis A, B, and C were negative, as were tests for antinuclear, antismooth muscle, and antimitochondrial antibodies. Serum ceruloplasmin, iron, TIBC, ferritin, and ␣1-antitrypsin were normal. There was normal hepatic echogenicity on ultrasound, and Doppler studies of the hepatic and portal veins were normal. The patient was instructed to stop troglitazone. Repeat liver chemistries in 4 wk revealed further elevation of serum aminotransferases, an increased total bilirubin and prothrombin time, and a decreased albumin (Fig. 1A). The patient had no arthralgias or skin rash and remained afebrile throughout the clinical course. She appeared jaundiced and
Troglitazone Hepatotoxicity
273
Figure 2. Histopathology of liver biopsies from two patients with troglitazone hepatotoxicity. A: Photomicrograph of biopsy from patient 1 showing zone 3 necrosis, characterized by hemorrhage, hepatocyte dropout, and a mild mononuclear cell infiltrate. A normal triad (t) is also visible (hematoxylin and eosin, ⫻100). B: Photomicrograph of biopsy from patient 2 showing cell swelling and some micro- and macrovesicular steatosis in zone 3 (lower left). The portal triad is relatively normal. Insert shows one of several scattered granulomas, which contains some lymphocytes and an eosinophil (e) (hematoxylin and eosin, ⫻100).
icteric, without other stigmata of liver disease, and the right upper quadrant of the abdomen was mildly tender. The patient said that, in error, she had continued taking troglitazone since her previous visit. Urgent liver biopsy showed zone 3 necrosis, hemorrhage, and hepatocyte dropout, with occasional foci of bridging necrosis to triads. In zones 1 and 2 there were active cell injury, as evidenced by acidophilic bodies, cell swelling, and lobular disarray. Macrovesicular steatosis was minimal, and microvesicular steatosis was not seen. Cholestasis was present, but not prominent. Kupffer cells were enlarged and appeared activated. Occasional hepatocyte mitoses were present. A mononuclear cell infiltrate was present diffusely, but concentrated at the edges of the areas of zone 3 necrosis. The triads and their structures were unremarkable. A trichrome stain showed early pericentral and bridging fibrosis (Fig. 2A). Periodic Acid-Schiff stain with diastase digestion showed PAS-positive debris in macrophages. Iron stain was negative.
274
Kohlroser et al.
The histology reflected hepatocellular injury that was severe, with zonal and bridging necrosis. The uniform pattern of necrosis and inflammation throughout the biopsy suggested that injury had occurred at one time; there was no evidence for underlying fibrosis or other disease processes. The mixture of necrosis, inflammation, and fibrosis suggested relatively recent, perhaps resolving, injury. Again, the patient was instructed to stop troglitazone and was given oral vitamin K supplementation. Eighteen days after stopping troglitazone the patient’s serum aminotransferases and bilirubin had decreased (Fig. 1A), and prothrombin time had become normal. She remained asymptomatic. Patient 2 A 62-year-old white man presented to our emergency room after 3 h of nausea, vomiting, dizziness, and palpitations. He complained of general malaise for the previous week. Two months earlier he had begun taking troglitazone 400 mg QD for type 2 diabetes mellitus. The patient also had a history of benign prostatic hypertrophy, elevated serum cholesterol, and systemic arterial hypertension. His other medications were simvastatin, 10 mg QD, metoprolol succinate extended release tablets, 50 mg QD, glyburide, 5 mg b.i.d., and aspirin 325 mg QD, all of which he had been taking for ⱖ2 yr. He was found to be in rapid atrial fibrillation, which responded to diltiazem HCL i.v. with return to normal sinus rhythm. Oral diltiazem was not prescribed. He rarely drank alcohol. There were no stigmata of liver disease; the abdominal examination was unremarkable. The patient had no skin rash, fever, arthralgia, or arthritis. Laboratory studies showed a serum creatinine of 1.5 mg/dl, hematocrit of 34%, and mean corpuscular volume of 81 m3. The white blood cell count was 6800/l with differential count as follows: 77% polys, 9% lymphocytes, 5% monocytes, and 1% eosinophils. The absolute eosinophil count was 68/l. Liver tests revealed a mixed hepatocellular-cholestatic picture with serum alkaline phosphatase of 253 U/L (reference range, 15– 85), ALT of 140 U/L (reference range, 10 – 40), AST of 65 U/L (reference range, 10 – 40), amylase of 27 U/L (reference range, 10 – 85), total bilirubin of 3.7 mg/dl (reference range, 0.3–1.2), and an albumin of 3.3 g/dl (reference range, 3.5–5.5) (Fig. 1B). The troglitazone and simvastatin were stopped on admission. Abdominal ultrasound demonstrated mild splenomegaly and a diffuse increase in echogenicity of the liver consistent with fatty infiltration, but no evidence of intra- or extrahepatic bile duct dilation and no gallstones. An ERCP showed no abnormalities of the bile ducts. A liver biopsy, done on hospital day 8, showed multiple nonnecrotizing granulomas, both in triads and lobules. The granulomas were composed of epithelioid histiocytes, multinucleated histiocytes, and white blood cells including many eosinophils. Necrosis was not present. No polarizable foreign material was detectable. A stain for acid fast bacilli and methenamine silver stain for fungi were negative. Some
AJG – Vol. 95, No. 1, 2000
of the histiocytic giant cells contained large lipid vacuoles, suggesting that these were lipid granulomas; otherwise, the hepatic architecture was intact. Inflammation was minimal and concentrated around the granulomas. A cell injury pattern was present, characterized by cell swelling and frequent binucleated cells. Steatosis, cholestasis, Mallory’s hyalin, and acidophilic bodies were not seen (Fig. 2B). The abnormal liver tests improved by day 8 and normalized completely by 4 months. He was restarted on all his previous medications except for the simvastatin and troglitazone. During the ensuing 9 months, liver chemistries continued to be normal. Evaluations for other liver disease showed negative serological tests for HBs Ag, anti-HBs, anti-HBc, anti-HCV, anti-HAV, antinuclear, and anti– smooth muscle antibodies, and normal values for serum ceruloplasmin, ␣1 antitrypsin, iron, TIBC, and ferritin.
REVIEW OF ADVERSE EVENTS REPORTED TO THE US FDA AND/OR RECENTLY PUBLISHED Under the terms of the Freedom of Information Act, we requested that the US Food and Drug Administration (FDA) search the Spontaneous Reporting System, an unverified voluntary database of adverse drug reactions. The FDA carried out a computer-assisted search of all MedWatch reports pertaining to troglitazone and provided a summary of 453 adverse reaction reports pertaining to troglitazone, submitted between March and October 1997. We searched this summary and requested copies of the original MedWatch forms for the 46 cases that we considered suspicious for hepatitis. If terms such as liver function abnormalities, hepatitis, hyperbilirubinemia, liver failure, and others were noted in the outline, the original MedWatch form was obtained. Of the 46 MedWatch reports, 45 dealt with presumed troglitazone-induced hepatitis. In the remaining case, a cholestatic hepatic enzyme profile was the result of common bile duct stones. Of the reports, 39 came from the United States and seven from foreign countries; 35 were submitted by the manufacturer, seven by medical professionals, and four by other individuals. A total of 27 MedWatch forms contained incomplete information. Results are summarized in Table 1. Striking results include the greater than 2:1 female:male ratio, the marked variability in cumulative drug dose (1,200 –78,000 mg), and duration of therapy (6 –195 days). Most patients had predominantly hepatocellular or mixed hepatocellular-cholestatic-type injury, and severity was substantial, with one man dying, one being evaluated for liver transplantation, and one of two women evaluated for liver transplantation receiving a new liver. It seems that women may be more susceptible; however, it may also be that more women than men have taken troglitazone. It is also likely that, among cases reported to the FDA, there is a greater than average severity.
AJG – January, 2000
Troglitazone Hepatotoxicity
275
Table 1. Summary of Adverse Events Related to Troglitazone Reported to the US Food and Drug Administration (Mar–Oct, 1997)
N Age, yr mean (range) Indication for starting troglitazone DM type unspecified DM type I DM type 2 Unknown Dosage (mg/day) unknown 200 mg 400 mg 600 mg Mean cumulative dose (mg) (range) Time from start of troglitazone to recognition of liver injury (days), mean (range) Peak serum level, mean (range) AST (U/L) ALT (U/L) Alkaline phosphatase (U/L) Total bilirubin (mg/dl) Outcome Complete recovery Evaluation for OLT OLT CBD Stones Unknown
Male
Female
Sex Not Specified
12 (26%) 61.6 (46–78)
27 (59%) 59.6 (33–72)
7 (15%) 60.3 (33–78)
2 1 8 1 1 4 4 3 35,533 (3,800–72,000) 88 (10–180)
6 1 14 6 7 3 16 1 25,371 (1,200–78,000) 63 (6–195)
1 0 0 6 6 1 0 0
973 (107–1836) 985 (257–2000) 588 (476–700) 7.44 (1.5–15.9)
766 (89–2000) 829 (89–2040) 329 (151–693) 9.41 (1.45–29)
3 1*
7 2† 1
8‡§
17
3 1† 1 2
OLT ⫽ orthotopic liver transplantation. * Patient died before transplant. † Complete recovery without transplant. ‡ One patient had bridging necrosis on liver biopsy, final outcome not reported. § One patient underwent plasmapheresis with initial improvement, final outcome not reported.
DISCUSSION From the patients described in this article (see Fig. 1 and 2) and our review of adverse events linked to troglitazone recently published and/or reported to the US FDA (Table 1), there can be no doubt that, in some patients, troglitazone causes severe hepatotoxicity. Furthermore, it does so “silently”: i.e., it does not cause a hypersensitivity or allergictype drug reaction, which would be more likely to alert patients or their physicians to drug-induced adverse effects. Rather, it causes liver injury that may be quite severe, as in our first patient, even in the absence of jaundice or any other major symptoms or signs of disease. Our second patient is, to our knowledge, the first in whom hepatitis with granulomas was associated with troglitazone use. Lipogranuloma formation is a reactive change reflective of nonspecific damage to lipid-laden hepatocytes. Other causes of granulomatous disease were excluded on clinical and histopathological grounds, and his rapid biochemical improvement after cessation of troglitazone makes the drug the most likely cause. We do not think that this is a case of classical granulomatous hepatitis. However, in view of the granuloma with eosinophils being present in the liver biopsy of the second case, the possibility of an allergic or immune type hypersensitivity cannot be excluded.
The interval between initiation of troglitazone therapy and recognition of significant hepatotoxicity has varied widely (Table 1). In many patients, however, including ours, frequent serum aminotransferase measurements were not performed. Thus, the time of onset of hepatotoxicity may have been earlier. Seen in this light, the current recommendations of the drug manufacturer, approved by the FDA, seem reasonable (3). These recommendations are that, in patients to receive troglitazone, serum liver chemistries (especially serum ALT) should be measured at baseline, monthly for the first 8 months, every other month for the next 4 months, and “periodically” thereafter. Troglitazone therapy should not be initiated in patients with a baseline ALT ⬎1.5 times the upper limit of normal (ULN) (3). The prevalence of increased serum ALT levels at baseline in patients for whom troglitazone is indicated will be higher than for healthy controls because of increased prevalence of fatty livers and/or steatohepatitis in the former. The manufacturer also recommends following weekly “liver function tests” in those patients with a serum ALT between 1.5 and 2 times ULN until return to baseline. We recommend that those patients with a persistent serum ALT 2 to 3 times ULN should either discontinue troglitazone or undergo liver biopsy. If appreciable necroinflammation, fibrosis, or lipo-
276
Kohlroser et al.
granulomas are found, use of the drug should be stopped. Those with mild changes in hepatic histopathology might continue, but only with close and frequent follow-up and after full disclosure of possible risks of progressive and perhaps fatal hepatic injury. We agree with the recommendation to stop therapy if the serum ALT is ⬎3 ULN. The proximate cause of troglitazone hepatotoxicity is not yet known, nor have clinical observations thus far revealed any clear ways to predict which patients are at increased risk for development of such toxicity (7, 8). Troglitazone is ⬎99% albumin-bound in the blood, has a long plasma half-life (16 –34 h), and is taken up well by the liver where it is metabolized chiefly by conjugation with sulfate and glucuronic acid. A small fraction is metabolized, probably by cytochrome(s) P-450, to a quinone, which then undergoes further phase II conjugation reactions, chiefly sulfation (9, 10). The form(s) of human cytochrome(s) P-450 that carry out metabolism of troglitazone are unknown. In earlier experimental studies, several expressed forms (1A1, 1A2, 2A6, 2B6, 2D6, 2E1, and 3A4) did not produce detectable metabolites. However, there is recent evidence for oxidation of troglitazone to a quinose-type metabolite, catalyzed by cytochromes P-450 2C8 and 3A4 in microsomes from human livers (11). Perhaps, patients at risk for troglitazone hepatotoxicity have a polymorphism in cytochrome P-450 or other metabolizing enzyme expression that produces more of a toxic, highly reactive intermediate. The current marketed form of troglitazone is an equal mixture of four stereoisomers, as the molecule contains two asymmetric chiral sites. It is not yet known whether, by separation of these isomers, it will be possible to dissociate the desirable from the undesirable effects of the drug. However, given the substantial benefits, but also the very real risks, of troglitazone, such effors seem to be worthy of effort.
ACKNOWLEDGMENTS We thank Susan Thayer-Reid and Alta Boover for typing the manuscript and Paula LeClair for helping us prepare Figure 1.
AJG – Vol. 95, No. 1, 2000
This work was supported by a grant from National Institutes of Health (NIH) (DK38825 to HLB) and by the Research and Education Fund of the Division of Digestive Disease and Nutrition, UMASS Memorial Health Care. The opinions expressed herein are those of the authors; they do not necessarily reflect the official views of the United States Public Health Service, the NIH, or the FDA. Reprint requests and correspondence: Herbert L. Bonkovsky, M.D., UMASS/Memorial Health Care, 55 Lake Avenue North, Worcester, MA 01655. Received Sep. 8, 1998; accepted Dec. 14, 1998.
REFERENCES 1. Henry RR. Thiazolidinediones. Endocrinol Metab Clin North Am 1997;26:553–73. 2. Spencer CM, Markham A. Troglitazone. Drugs 1997;54:89 – 102. 3. Prescribing Information for Rezulin. July 1998. Morris Plains, NJ: Parke-Davis. 4. US Food and Drug Administration. Patient testing and labeling strengthened for Rezulin. FDA Talk Paper: T97-61. Dec. 1, 1997. Rockville, MD: US Food and Drug Administration. 5. Gitlin N, Neil JL, Spurr CL, et al. Two cases of severe clinical and histologic hepatotoxicity associated with troglitazone. Ann Intern Med. 1998;129:36 – 8. 6. Neuschwander-Tetri BA, Isley WL, Oki JC, et al. Troglitazone-induced hepatic failure leading to liver transplantation: A case report. Ann Intern Med 1998;129:38 – 41. 7. Watkins PB, Whitcomb RW. Hepatic dysfunction associated with troglitazone. N Engl J Med 1998;388:916 –7. 8. Imura H. Troglitazone. N Engl J Med 1998;338:808 –9 (editorial). 9. Izumi I, Hosiyama K, Enomoto S, et al. Pharmacokinetics of troglitazone, an antidiabetic agent: Prediction of in vivo stereoselective sulfation and glucuronidation from in vitro data. J Pharmacol Exptl Therap 1997;280:1392–1400. 10. Kawai K, Kawasaki-Tokui Y, Odaka T, et al. Disposition and metabolism of the new oral antidiabetic drug troglitazone in rats, mice and dogs. Arzneimittel-Forschung 1997;47:356-68. 11. Yamazaki H, Shibata A, Suzuki M, et al. Oxidation of troglitazone to a quinone-type metabolite catalyzed by cytochrome P-450 2C8 and P-450 3A4 in human liver microsomes. Drug Metab Dispos 1999;27:1260 – 6.
Vol. 95, No. 1, 2000 ISSN 0002-9270/00/$20.00 PII S0002-9270(99)00766-2
CASE REPORTS
Hepatotoxicity Due to Troglitazone: Report of Two Cases and Review of Adverse Events Reported to the United States Food and Drug Administration James Kohlroser, D.O., Jijy Mathai, M.D., James Reichheld, M.D., Barbara F. Banner, M.D., and Herbert L. Bonkovsky, M.D. Departments of Medicine, Pathology, Biochemistry and Molecular Biology, and The Liver, Biliary, Pancreatic Center of UMASS Memorial Health Care and the University of Massachusetts Medical School, Worcester, Machusetts
ABSTRACT Two patients (a 48-year-old woman and a 62-year-old man) developed clinical and laboratory signs of hepatotoxicity due to troglitazone (Rezulin), a thiazolidinedione used in treatment of diabetes mellitus. There was no clear clinical evidence of drug allergy, although the woman experienced colitis before the onset of recognized hepatotoxicity. Liver biopsies showed bridging necrosis and fibrosis in the woman and hepatitis with granuloma formation in the man. The abnormalities in liver chemistries resolved promptly upon cessation of the drug. Cases involving 46 patients reported to the United States Food and Drug Administration are also reviewed. Troglitazone is a useful new oral antihyperglycemic agent, but in about 1.9% of patients hepatotoxicity has occurred, which may be severe and even fatal. Frequent monitoring of serum liver chemistries in patients taking the drug is essential. (Am J Gastroenterol 2000;95: 272–276. © 2000 by Am. Coll. of Gastroenterology)
INTRODUCTION Troglitazone is a thiazolidinedione with insulin-sensitizing activities when administered to humans or animals with type 2 (noninsulin-dependent) diabetes mellitus. It has been shown to have a number of desirable metabolic effects on glucose and lipid metabolism, and it has been approved for clinical use in Europe, Japan, and the United States (1, 2). A major adverse effect of troglitazone is the development of hepatotoxicity. In early clinical trials, elevations of serum aminotransferases (⬎3 times the upper limit of normal) occurred in 48 of 2510 (1.9%) subjects receiving troglitazone (3). In December 1997 and again in August 1998, the United States Food and Drug Administration (FDA) issued stronger warnings and recommendations after receiving reports of ⬎100 cases of liver damage, including liver failure requiring transplantation in three patients and death in another patient treated with troglitazone (4). These events must be balanced against the beneficial effects of the drug on glucose and lipid metabolism and interpreted in the light
of the large numbers of patients who have been, and are being treated with it (⬎800,000 in the US and Japan). Concerns about the hepatotoxicity of troglitazone led the Medicines Control Agency of the United Kingdom to request voluntary withdrawal of the drug from the UK, effective December 1, 1997 (5). This request was honored by the pharmaceutical companies involved. In this article, we describe two previously unreported, carefully studied patients with troglitazone hepatotoxicity, comparing their features with other reported patients, including review of 46 cases of hepatotoxicity reported to the FDA, information on which we obtained under the Freedom of Information Act. Three case reports have recently been published reporting troglitazone induced hepatotoxicity. All three were reported to the FDA; two of the patients improved with discontinuation of troglitazone and supportive care, whereas the third patient developed hepatic encephalopathy and worsening liver failure requiring liver transplantation (5, 6) Because all identifying information is removed from the Medwatch reports before being released under the Freedom of Information Act, we cannot ascertain whether these cases are included in the reports that we obtained from the FDA.
CASE REPORTS Patient 1 A 48-year-old Hispanic woman developed jaundice 8 months after starting troglitazone for the treatment of diabetes mellitus type 2. Her initial dose was 200 mg b.i.d., which was increased to 400 mg b.i.d. after the first 3 months. Notably, her serum ALT had been normal 3 yr before beginning troglitazone (34 U/L), but was mildly elevated (44 U/L) after 1 month of therapy (Fig. 1A). She had no known drug allergies. She presented with diarrhea 5 months after beginning troglitazone and was found to have mild colitis by colonoscopy. Stool cultures for enteric pathogens, Yersinia, and Campylobacter were negative, as were examinations for ova
AJG – January, 2000
Figure 1. Summary of liver chemistries of two patients with troglitazone hepatotoxicity. A: Patient 1. B: Patient 2.
and parasites. Her diarrhea resolved without further intervention. She denied the use of other prescription or overthe-counter medications and rarely used alcohol. On examination, she was without lymphadenopathy; there were no stigmata of liver disease and there was no abdominal tenderness, mass, or hepatosplenomegaly. Liver chemistries revealed elevated serum aminotransferases, ALT 653 U/L, AST 413 U/L (normal ⱕ40 U/L), and total bilirubin 1.5 mg/dl (normal ⱕ1.2 mg/dl) (Fig. 1A). The complete blood count was normal with white blood cell count of 6700/l and a manual differential of 50% polys (polymorphonuclear leukocytes), 35% lymphocytes, 13% monocytes, 1.8% eosinophils, and 0.8% basophils. Serologies for hepatitis A, B, and C were negative, as were tests for antinuclear, antismooth muscle, and antimitochondrial antibodies. Serum ceruloplasmin, iron, TIBC, ferritin, and ␣1-antitrypsin were normal. There was normal hepatic echogenicity on ultrasound, and Doppler studies of the hepatic and portal veins were normal. The patient was instructed to stop troglitazone. Repeat liver chemistries in 4 wk revealed further elevation of serum aminotransferases, an increased total bilirubin and prothrombin time, and a decreased albumin (Fig. 1A). The patient had no arthralgias or skin rash and remained afebrile throughout the clinical course. She appeared jaundiced and
Troglitazone Hepatotoxicity
273
Figure 2. Histopathology of liver biopsies from two patients with troglitazone hepatotoxicity. A: Photomicrograph of biopsy from patient 1 showing zone 3 necrosis, characterized by hemorrhage, hepatocyte dropout, and a mild mononuclear cell infiltrate. A normal triad (t) is also visible (hematoxylin and eosin, ⫻100). B: Photomicrograph of biopsy from patient 2 showing cell swelling and some micro- and macrovesicular steatosis in zone 3 (lower left). The portal triad is relatively normal. Insert shows one of several scattered granulomas, which contains some lymphocytes and an eosinophil (e) (hematoxylin and eosin, ⫻100).
icteric, without other stigmata of liver disease, and the right upper quadrant of the abdomen was mildly tender. The patient said that, in error, she had continued taking troglitazone since her previous visit. Urgent liver biopsy showed zone 3 necrosis, hemorrhage, and hepatocyte dropout, with occasional foci of bridging necrosis to triads. In zones 1 and 2 there were active cell injury, as evidenced by acidophilic bodies, cell swelling, and lobular disarray. Macrovesicular steatosis was minimal, and microvesicular steatosis was not seen. Cholestasis was present, but not prominent. Kupffer cells were enlarged and appeared activated. Occasional hepatocyte mitoses were present. A mononuclear cell infiltrate was present diffusely, but concentrated at the edges of the areas of zone 3 necrosis. The triads and their structures were unremarkable. A trichrome stain showed early pericentral and bridging fibrosis (Fig. 2A). Periodic Acid-Schiff stain with diastase digestion showed PAS-positive debris in macrophages. Iron stain was negative.
274
Kohlroser et al.
The histology reflected hepatocellular injury that was severe, with zonal and bridging necrosis. The uniform pattern of necrosis and inflammation throughout the biopsy suggested that injury had occurred at one time; there was no evidence for underlying fibrosis or other disease processes. The mixture of necrosis, inflammation, and fibrosis suggested relatively recent, perhaps resolving, injury. Again, the patient was instructed to stop troglitazone and was given oral vitamin K supplementation. Eighteen days after stopping troglitazone the patient’s serum aminotransferases and bilirubin had decreased (Fig. 1A), and prothrombin time had become normal. She remained asymptomatic. Patient 2 A 62-year-old white man presented to our emergency room after 3 h of nausea, vomiting, dizziness, and palpitations. He complained of general malaise for the previous week. Two months earlier he had begun taking troglitazone 400 mg QD for type 2 diabetes mellitus. The patient also had a history of benign prostatic hypertrophy, elevated serum cholesterol, and systemic arterial hypertension. His other medications were simvastatin, 10 mg QD, metoprolol succinate extended release tablets, 50 mg QD, glyburide, 5 mg b.i.d., and aspirin 325 mg QD, all of which he had been taking for ⱖ2 yr. He was found to be in rapid atrial fibrillation, which responded to diltiazem HCL i.v. with return to normal sinus rhythm. Oral diltiazem was not prescribed. He rarely drank alcohol. There were no stigmata of liver disease; the abdominal examination was unremarkable. The patient had no skin rash, fever, arthralgia, or arthritis. Laboratory studies showed a serum creatinine of 1.5 mg/dl, hematocrit of 34%, and mean corpuscular volume of 81 m3. The white blood cell count was 6800/l with differential count as follows: 77% polys, 9% lymphocytes, 5% monocytes, and 1% eosinophils. The absolute eosinophil count was 68/l. Liver tests revealed a mixed hepatocellular-cholestatic picture with serum alkaline phosphatase of 253 U/L (reference range, 15– 85), ALT of 140 U/L (reference range, 10 – 40), AST of 65 U/L (reference range, 10 – 40), amylase of 27 U/L (reference range, 10 – 85), total bilirubin of 3.7 mg/dl (reference range, 0.3–1.2), and an albumin of 3.3 g/dl (reference range, 3.5–5.5) (Fig. 1B). The troglitazone and simvastatin were stopped on admission. Abdominal ultrasound demonstrated mild splenomegaly and a diffuse increase in echogenicity of the liver consistent with fatty infiltration, but no evidence of intra- or extrahepatic bile duct dilation and no gallstones. An ERCP showed no abnormalities of the bile ducts. A liver biopsy, done on hospital day 8, showed multiple nonnecrotizing granulomas, both in triads and lobules. The granulomas were composed of epithelioid histiocytes, multinucleated histiocytes, and white blood cells including many eosinophils. Necrosis was not present. No polarizable foreign material was detectable. A stain for acid fast bacilli and methenamine silver stain for fungi were negative. Some
AJG – Vol. 95, No. 1, 2000
of the histiocytic giant cells contained large lipid vacuoles, suggesting that these were lipid granulomas; otherwise, the hepatic architecture was intact. Inflammation was minimal and concentrated around the granulomas. A cell injury pattern was present, characterized by cell swelling and frequent binucleated cells. Steatosis, cholestasis, Mallory’s hyalin, and acidophilic bodies were not seen (Fig. 2B). The abnormal liver tests improved by day 8 and normalized completely by 4 months. He was restarted on all his previous medications except for the simvastatin and troglitazone. During the ensuing 9 months, liver chemistries continued to be normal. Evaluations for other liver disease showed negative serological tests for HBs Ag, anti-HBs, anti-HBc, anti-HCV, anti-HAV, antinuclear, and anti– smooth muscle antibodies, and normal values for serum ceruloplasmin, ␣1 antitrypsin, iron, TIBC, and ferritin.
REVIEW OF ADVERSE EVENTS REPORTED TO THE US FDA AND/OR RECENTLY PUBLISHED Under the terms of the Freedom of Information Act, we requested that the US Food and Drug Administration (FDA) search the Spontaneous Reporting System, an unverified voluntary database of adverse drug reactions. The FDA carried out a computer-assisted search of all MedWatch reports pertaining to troglitazone and provided a summary of 453 adverse reaction reports pertaining to troglitazone, submitted between March and October 1997. We searched this summary and requested copies of the original MedWatch forms for the 46 cases that we considered suspicious for hepatitis. If terms such as liver function abnormalities, hepatitis, hyperbilirubinemia, liver failure, and others were noted in the outline, the original MedWatch form was obtained. Of the 46 MedWatch reports, 45 dealt with presumed troglitazone-induced hepatitis. In the remaining case, a cholestatic hepatic enzyme profile was the result of common bile duct stones. Of the reports, 39 came from the United States and seven from foreign countries; 35 were submitted by the manufacturer, seven by medical professionals, and four by other individuals. A total of 27 MedWatch forms contained incomplete information. Results are summarized in Table 1. Striking results include the greater than 2:1 female:male ratio, the marked variability in cumulative drug dose (1,200 –78,000 mg), and duration of therapy (6 –195 days). Most patients had predominantly hepatocellular or mixed hepatocellular-cholestatic-type injury, and severity was substantial, with one man dying, one being evaluated for liver transplantation, and one of two women evaluated for liver transplantation receiving a new liver. It seems that women may be more susceptible; however, it may also be that more women than men have taken troglitazone. It is also likely that, among cases reported to the FDA, there is a greater than average severity.
AJG – January, 2000
Troglitazone Hepatotoxicity
275
Table 1. Summary of Adverse Events Related to Troglitazone Reported to the US Food and Drug Administration (Mar–Oct, 1997)
N Age, yr mean (range) Indication for starting troglitazone DM type unspecified DM type I DM type 2 Unknown Dosage (mg/day) unknown 200 mg 400 mg 600 mg Mean cumulative dose (mg) (range) Time from start of troglitazone to recognition of liver injury (days), mean (range) Peak serum level, mean (range) AST (U/L) ALT (U/L) Alkaline phosphatase (U/L) Total bilirubin (mg/dl) Outcome Complete recovery Evaluation for OLT OLT CBD Stones Unknown
Male
Female
Sex Not Specified
12 (26%) 61.6 (46–78)
27 (59%) 59.6 (33–72)
7 (15%) 60.3 (33–78)
2 1 8 1 1 4 4 3 35,533 (3,800–72,000) 88 (10–180)
6 1 14 6 7 3 16 1 25,371 (1,200–78,000) 63 (6–195)
1 0 0 6 6 1 0 0
973 (107–1836) 985 (257–2000) 588 (476–700) 7.44 (1.5–15.9)
766 (89–2000) 829 (89–2040) 329 (151–693) 9.41 (1.45–29)
3 1*
7 2† 1
8‡§
17
3 1† 1 2
OLT ⫽ orthotopic liver transplantation. * Patient died before transplant. † Complete recovery without transplant. ‡ One patient had bridging necrosis on liver biopsy, final outcome not reported. § One patient underwent plasmapheresis with initial improvement, final outcome not reported.
DISCUSSION From the patients described in this article (see Fig. 1 and 2) and our review of adverse events linked to troglitazone recently published and/or reported to the US FDA (Table 1), there can be no doubt that, in some patients, troglitazone causes severe hepatotoxicity. Furthermore, it does so “silently”: i.e., it does not cause a hypersensitivity or allergictype drug reaction, which would be more likely to alert patients or their physicians to drug-induced adverse effects. Rather, it causes liver injury that may be quite severe, as in our first patient, even in the absence of jaundice or any other major symptoms or signs of disease. Our second patient is, to our knowledge, the first in whom hepatitis with granulomas was associated with troglitazone use. Lipogranuloma formation is a reactive change reflective of nonspecific damage to lipid-laden hepatocytes. Other causes of granulomatous disease were excluded on clinical and histopathological grounds, and his rapid biochemical improvement after cessation of troglitazone makes the drug the most likely cause. We do not think that this is a case of classical granulomatous hepatitis. However, in view of the granuloma with eosinophils being present in the liver biopsy of the second case, the possibility of an allergic or immune type hypersensitivity cannot be excluded.
The interval between initiation of troglitazone therapy and recognition of significant hepatotoxicity has varied widely (Table 1). In many patients, however, including ours, frequent serum aminotransferase measurements were not performed. Thus, the time of onset of hepatotoxicity may have been earlier. Seen in this light, the current recommendations of the drug manufacturer, approved by the FDA, seem reasonable (3). These recommendations are that, in patients to receive troglitazone, serum liver chemistries (especially serum ALT) should be measured at baseline, monthly for the first 8 months, every other month for the next 4 months, and “periodically” thereafter. Troglitazone therapy should not be initiated in patients with a baseline ALT ⬎1.5 times the upper limit of normal (ULN) (3). The prevalence of increased serum ALT levels at baseline in patients for whom troglitazone is indicated will be higher than for healthy controls because of increased prevalence of fatty livers and/or steatohepatitis in the former. The manufacturer also recommends following weekly “liver function tests” in those patients with a serum ALT between 1.5 and 2 times ULN until return to baseline. We recommend that those patients with a persistent serum ALT 2 to 3 times ULN should either discontinue troglitazone or undergo liver biopsy. If appreciable necroinflammation, fibrosis, or lipo-
276
Kohlroser et al.
granulomas are found, use of the drug should be stopped. Those with mild changes in hepatic histopathology might continue, but only with close and frequent follow-up and after full disclosure of possible risks of progressive and perhaps fatal hepatic injury. We agree with the recommendation to stop therapy if the serum ALT is ⬎3 ULN. The proximate cause of troglitazone hepatotoxicity is not yet known, nor have clinical observations thus far revealed any clear ways to predict which patients are at increased risk for development of such toxicity (7, 8). Troglitazone is ⬎99% albumin-bound in the blood, has a long plasma half-life (16 –34 h), and is taken up well by the liver where it is metabolized chiefly by conjugation with sulfate and glucuronic acid. A small fraction is metabolized, probably by cytochrome(s) P-450, to a quinone, which then undergoes further phase II conjugation reactions, chiefly sulfation (9, 10). The form(s) of human cytochrome(s) P-450 that carry out metabolism of troglitazone are unknown. In earlier experimental studies, several expressed forms (1A1, 1A2, 2A6, 2B6, 2D6, 2E1, and 3A4) did not produce detectable metabolites. However, there is recent evidence for oxidation of troglitazone to a quinose-type metabolite, catalyzed by cytochromes P-450 2C8 and 3A4 in microsomes from human livers (11). Perhaps, patients at risk for troglitazone hepatotoxicity have a polymorphism in cytochrome P-450 or other metabolizing enzyme expression that produces more of a toxic, highly reactive intermediate. The current marketed form of troglitazone is an equal mixture of four stereoisomers, as the molecule contains two asymmetric chiral sites. It is not yet known whether, by separation of these isomers, it will be possible to dissociate the desirable from the undesirable effects of the drug. However, given the substantial benefits, but also the very real risks, of troglitazone, such effors seem to be worthy of effort.
ACKNOWLEDGMENTS We thank Susan Thayer-Reid and Alta Boover for typing the manuscript and Paula LeClair for helping us prepare Figure 1.
AJG – Vol. 95, No. 1, 2000
This work was supported by a grant from National Institutes of Health (NIH) (DK38825 to HLB) and by the Research and Education Fund of the Division of Digestive Disease and Nutrition, UMASS Memorial Health Care. The opinions expressed herein are those of the authors; they do not necessarily reflect the official views of the United States Public Health Service, the NIH, or the FDA. Reprint requests and correspondence: Herbert L. Bonkovsky, M.D., UMASS/Memorial Health Care, 55 Lake Avenue North, Worcester, MA 01655. Received Sep. 8, 1998; accepted Dec. 14, 1998.
REFERENCES 1. Henry RR. Thiazolidinediones. Endocrinol Metab Clin North Am 1997;26:553–73. 2. Spencer CM, Markham A. Troglitazone. Drugs 1997;54:89 – 102. 3. Prescribing Information for Rezulin. July 1998. Morris Plains, NJ: Parke-Davis. 4. US Food and Drug Administration. Patient testing and labeling strengthened for Rezulin. FDA Talk Paper: T97-61. Dec. 1, 1997. Rockville, MD: US Food and Drug Administration. 5. Gitlin N, Neil JL, Spurr CL, et al. Two cases of severe clinical and histologic hepatotoxicity associated with troglitazone. Ann Intern Med. 1998;129:36 – 8. 6. Neuschwander-Tetri BA, Isley WL, Oki JC, et al. Troglitazone-induced hepatic failure leading to liver transplantation: A case report. Ann Intern Med 1998;129:38 – 41. 7. Watkins PB, Whitcomb RW. Hepatic dysfunction associated with troglitazone. N Engl J Med 1998;388:916 –7. 8. Imura H. Troglitazone. N Engl J Med 1998;338:808 –9 (editorial). 9. Izumi I, Hosiyama K, Enomoto S, et al. Pharmacokinetics of troglitazone, an antidiabetic agent: Prediction of in vivo stereoselective sulfation and glucuronidation from in vitro data. J Pharmacol Exptl Therap 1997;280:1392–1400. 10. Kawai K, Kawasaki-Tokui Y, Odaka T, et al. Disposition and metabolism of the new oral antidiabetic drug troglitazone in rats, mice and dogs. Arzneimittel-Forschung 1997;47:356-68. 11. Yamazaki H, Shibata A, Suzuki M, et al. Oxidation of troglitazone to a quinone-type metabolite catalyzed by cytochrome P-450 2C8 and P-450 3A4 in human liver microsomes. Drug Metab Dispos 1999;27:1260 – 6.