Charles L. Spurr

Two Cases of Severe Clinical and Histologic Hepatotoxicity Associated with Troglitazone

Several oral antihyperglycemic agents are available for management of type 2 diabetes. These agents generally fall into the chemical or function groups sulfonylureas, biguanides, or -glucosidase inhibitors. Recently, the U.S. Food and Drug Administration approved troglitazone (Rezulin, Parke-Davis, Morris Plains, New Jersey), a thiazolidinedione agent that is unrelated to the aforementioned categories, as a novel oral antihyperglycemic agent. It decreases hepatic glucose output and increases insulin-dependent glucose metabolism in skeletal muscle [1, 2]. It is not an insulin secretagogue. We report two patients who had severe hepatotoxicity associated with troglitazone administration. Patient 1 A 44-year-old woman with a history of type 1 diabetes mellitus since 1978 presented with nausea, low-grade fever, and progressive jaundice. She had required insulin for diabetes control for 10 years. Comorbid conditions included essential hypertension, hypothyroidism, and obesity. The patient had started taking troglitazone 20 weeks before presentation. Initially, she received 200 mg daily for 4 weeks, then received 400 mg daily thereafter. Before starting therapy, the patients hepatic biochemical markers were normal. Other medications included lisinopril, 20 mg/d; levothyroxine sodium, 88 g/d; and insulin. The patient reported no history of exposure to hepatotoxins or alcohol ingestion. On examination, the patient was icteric, and her liver was tender and enlarged to 14 cm; she was hospitalized for evaluation. Results of serology for hepatitis A, B, and C and studies for cytomegalovirus and Epstein-Barr virus were negative. Immunoglobulin levels were normal, and the ratio of antinuclear antibodies was positive (1:80). Troglitazone therapy was discontinued, and other therapies were maintained. Table 1 summarizes the patient characteristics. Table 1. Patient Characteristics The patient had a protracted hospital course with prolonged cholestasis, a decrease in albumin level to 18 g/L, a prothrombin time of 14.8 seconds, and development of ascites. Because of concern that the patient was developing subacute hepatic necrosis, a liver biopsy was performed (Figure 1). Biopsy results showed diffuse necrosis of hepatocytes with bridging necrosis, regenerating hepatocytes, intracanalicular cholestasis, no substantial steatosis, and no cellular infiltrate. Early fibrosis was evident. The patient gradually improved over the subsequent 3 months: Her bilirubin level decreased to 171 mol/L, her albumin level increased to 27 g/L, and her prothrombin time returned to normal. Figure 1. Low-power view of portal tract showing interface hepatitis, mild inflammatory infiltrate, bile duct reduplication, and necrotic hepatocytes (hematoxylin-eosin stain; original magnification, x96). Patient 2 A 65-year-old woman presented with a 3-week history of nausea, anorexia, pruritus, and jaundice. She had had type 2 diabetes for 4 years. Comorbid conditions included obesity, essential hypertension, mild osteoarthritis, and a solitary renal calculus. She had started therapy with troglitazone, 400 mg/d, 6 weeks before presentation. Other medications included insulin; atenolol, 50 mg/d; and furosemide, 40 mg/d, for the past 2 years. The patient had no history of alcohol ingestion, liver disease, blood transfusion, or known exposure to hepatotoxins or potentially hepatotoxic chemicals. No information from the physical examination was available. Laboratory values at presentation were as follows: bilirubin level, 342 mol/L; alanine aminotransferase level, 19.42 kat/L; cholesterol, 12.9 mol/L; and prothrombin time, 12.6 seconds. Serum protein levels were normal (Table 1). Results of serology were negative for hepatitis A, B, and C and infectious mononucleosis. Results of an antinuclear antibody test and endoscopic retrograde cholangiopancreatography were negative. Treatment with troglitazone was discontinued, and other therapies were maintained. Eight weeks after presentation, she underwent liver biopsy. Histology revealed hepatocyte necrosis with bridging necrosis and fibrosis, a portal inflammatory infiltrate (mostly plasma cells and eosinophils) and minimal cholestasis. Three months after discontinuation of troglitazone therapy, the patients bilirubin level was 78.66 mol/L and her alanine aminotransferase level was 5.48 kat/L. Discussion Drug hepatotoxicity is usually attributable to an intrinsic or an idiosyncratic hepatotoxin [3, 4]. The former is a dose-dependent, reproducible mechanism that is common, well understood, and preventable. The latter mechanism is not dose-dependent, is rare, is not reproducible, and usually reflects host hypersensitivity to an intermediate metabolite of the parent molecule. Troglitazone is rapidly absorbed after oral administration, reaching maximal plasma concentration within 2 to 3 hours. Concurrent administration with food accelerates the rate of absorption. Troglitazone is highly bound to albumin and is metabolized in the liver, mostly to a sulfate conjugate. It induces cytochrome P4503A4 and does not seem to inhibit cytochrome P450 enzymes. The induction of cytochrome P4503A4 may result in accelerated metabolism of other medications that make use of this enzyme. No other suspected drug interactions were clinically evident in the two patients discussed here. In studies of troglitazone in patients with chronic liver disease (Child-Pugh grade B or C), elevated levels of sulphate conjugates were noted without any adverse increase or change in hepatic biochemistry[5]. In clinical studies in North America, a total of 2510 patients received troglitazone; 20 (2.2%) of these patients had increases in the alanine aminotransferase level that were greater than threefold above normal, and 2 of these patients developed reversible jaundice (Parke-Davis, Morris Plains, New Jersey. Data on file). We describe two patients with severe hepatotoxicity caused by troglitazone therapy after 138 days and 35 days of exposure to standard dosages of this medication. Neither patient had preexisting hepatic or renal disease. Patient 1 was also receiving lisinopril, which can cause cholestatic hepatitis; however, the patient had been receiving this agent for more than 1 year. Histologic and biochemical findings in both patients showed a predominant acute hepatocellular necrosis with bridging necrosis and fibrosis with some evidence of cholestasis. Hepatotoxicity in this case may reflect an idiosyncratic mechanism caused by a hypersensitivity to a troglitazone metabolite. It is anticipated that further evidence with the use of troglitazone will reveal the extent and frequency of this hepatotoxicity. Until such information is available, physicians and patients must be aware of this potential adverse association. We recommend close monitoring of biochemical markers in persons receiving troglitazone. The manufacturer of troglitazone and the U.S. Food and Drug Administration were notified of these cases. Distribution of the medication by a licensee was halted in the United Kingdom in November 1997 on the basis of reports of hepatic dysfunction. The manufacturer of troglitazone altered its prescribing information on 3 November 1997 to specify a need for monitoring liver function in patients taking this drug. Dr. Julie: 15225 Shady Grove Road, #207, Rockville, MD 20852. Dr. Spurr: 820 St. Sebastian Way, #5B, Augusta, GA 30901. Dr. Juarbe: 19241 Montgomery Village Avenue, Suite E23, Montgomery Village, MD 20886.

Interrupted versus Continuous Chemotherapy in Patients with Metastatic Breast Cancer

BACKGROUND Chemotherapy for metastatic breast cancer is palliative, and the optimal duration of therapy is unknown. We designed a trial to determine whether continuous treatment is superior to stopping treatment after a brief induction period and resuming treatment when the disease progresses. METHODS We treated 250 women with metastatic breast cancer with six courses of cyclophosphamide, doxorubicin, and fluorouracil given every three weeks. At the completion of this induction period, women whose disease either regressed or remained stable were randomly assigned to receive either continued treatment with cyclophosphamide, methotrexate, and fluorouracil (maintenance therapy) or no further treatment (observation) followed by treatment with cyclophosphamide, methotrexate, and fluorouracil when disease progression became evident (reinduction). RESULTS The combined rate of complete and partial responses after initial therapy was 30 percent (71 of 233 patients who could be evaluated; 95 percent confidence interval, 25 percent to 37 percent). In another 42 percent (98 patients), the disease remained stable. A total of 145 patients were randomized. Seventy-one were randomly assigned to the maintenance-therapy group, and 74 to the observation group. The median time to progression was 9.4 months for patients in the maintenance-therapy group and 3.2 months for patients in the observation group (P less than 0.001). After reinduction therapy, the median time to progression was 3.5 months. The median length of survival from the time of initial therapy was 14.8 months for all 250 patients; it was 21.1 months for the 71 patients in the maintenance-therapy group and 19.6 months for the 74 patients in the observation group (P = 0.67). Maintenance therapy was the most important determinant of the time before progression (P less than 0.001), but it was not associated with prolonged survival. The changes in performance status were similar in the patients in both groups, but nausea, vomiting, and mucositis were significantly more frequent in the maintenance-therapy group. CONCLUSIONS In patients with breast cancer who received induction chemotherapy for 18 weeks, subsequent continuous chemotherapy was associated with a significant prolongation of the time before progression as compared with those receiving no further therapy; overall survival, however, was not significantly different in the two groups.

Complete Deficiency of Leukocyte Glucose-6-Phosphate Dehydrogenase with Defective Bactericidal Activity

A 52 yr old Caucasian female (F. E.) had hemolytic anemia, a leukemoid reaction, and fatal sepsis due to Escherichia coli. Her leukocytes ingested bacteria normally but did not kill catalase positive Staphylococcus aureus, Escherichia coli, and Serratia marcescens. An H(2)O(2)-producing bacterium, Streptococcus faecalis, was killed normally. Granule myeloperoxidase, acid and alkaline phosphatase, and beta glucuronidase activities were normal, and these enzymes shifted normally to the phagocyte vacuole (light and electron microscopy). Intravacuolar reduction of nitroblue tetrazolium did not occur. Moreover, only minimal quantities of H(2)O(2) were generated, and the hexose monophosphate shunt (HMPS) was not stimulated during phagocytosis. These observations suggested the diagnosis of chronic granulomatous disease. However, in contrast to control and chronic granulomatous disease leukocytes, glucose-6-phosphate dehydrogenase activity was completely absent in F. E. leukocytes whereas NADH oxidase and NADPH oxidase activities were both normal. Unlike chronic granulomatous disease, methylene blue did not stimulate the hexose monophosphate shunt in F. E. cells. Thus, F. E. and chronic granulomatous disease leukocytes appear to share certain metabolic and bactericidal defects, but the metabolic basis of the abnormality differs. Chronic granulomatous disease cells lack oxidase activity which produces H(2)O(2); F. E. cells had normal levels of oxidase activity but failed to produce NADPH due to complete glucose-6-phosphate dehydrogenase deficiency. These data indicate that a complete absence of leukocyte glucose-6-phosphate dehydrogenase with defective hexose monophosphate shunt activity is associated with low H(2)O(2) production and inadequate bactericidal activity, and further suggest an important role for NADPH in the production of H(2)O(2) in human granulocytes.

Written informed consent in patients with breast cancer

One hundred breast cancer patients, 35 adjuvant and 65 advanced, were interviewed 0–24 months after the start of chemotherapy to assess their knowledge and perceptions of the purposes, risks and benefits of treatment. Prior to therapy, all had been given verbal explanations and had signed informed consent forms explicitly detailing drugs, objectives and possible adverse effects of therapy. Seventeen percent of adjuvant and 29% of advanced patients were unable to name any of their drugs. While most patients recognized distressing side effects such as nausea and hair loss, less than 50% were aware of the potentially lethal complications of infection and bleeding. While the purpose of adjuvant therapy was cure, only 29% of the adjuvant patients were aware of this. In contrast, 35% of the advanced patients incorrectly stated that they were told their therapy was potentially curative. Explanations given by a nurse in addition to a physician were better understood than those given by a physician alone. We conclude that, in spite of intensive efforts at improving informed consent procedures, current results are unsatisfactory.

Adriamycin versus methotrexate in five-drug combination chemotherapy for advanced breast cancer: a randomized trial.

Adriamycin is of noteworthy efficacy in the treatment of metastatic breast cancer. Its role in combination regimens is under investigation. One hundred seventy‐five women with advanced breast cancer were entered into a prospectively randomized trial comparing two five‐drug regimens. Regimen CMFVP consisted of cyclophosphamide (C), methotrexate (M), 5‐fluorouracil (F), vincristine (V), and prednisone (P). Regimen CAFVP was identical but substituted Adriamycin (A) for methotrexate. Twenty‐seven patients were disqualified; 148 were evaluable. With CMFVP the complete response rate (CR) was 11%, and the partial response rate (PR) was 46%; with CAFVP, CR was 13% and PR was 45%. Duration of response tended to be slightly longer for patients on the Adriamycin arm. The median survival for CR and PR patients with CMFVP was 20.2 months, which was shorter (p =.07) than the 33 month median survival with CAFVP. Although statistical significance was not reached at the 5% level, the increased survival of responders on the Adriamycin regimen supports the data of other studies which suggest that first line combination chemotherapy in advanced breast cancer should include Adriamycin. Cancer 42:2141–2148, 1978.

Comparative study of cytosine arabinoside therapy alone and combined with thioguanine, mercaptopurine, or daunorubicin in acute myelocytic leukemia.

Three hundred twenty‐six patients with acute myelocytic leukemia were randomly and prospectively assigned to four therapeutic regimens: cytosine arabinoside either alone or in combination with daunorubicin, 6‐mercaptopurine, or 6‐thioguanine. The results in 231 qualified previously untreated patients were analyzed. The combination treatments produced a significantly greater frequency of complete or partial remission than single drug therapy. Treatment with cytosine arabinoside and thioguanine led to 48% ageadjusted complete and partial responses. The median survival from diagnosis of all 66 evaluable patients treated with these two drugs was 18 weeks, while the median survival for those who responded to this combination was 15 months.

A randomized comparative trial of chemotherapy and irradiation therapy for stage ii breast cancer

One hundred fifty‐eight evaluable patients with Stage II carcinoma of the breast with positive lymph nodes were treated either with adjuvant chemotherapy or irradiation therapy followed by chemotherapy. Patients were randomized to test the effectiveness of L‐phenylalanine mustard (L‐PAM) with and without postoperative irradiation therapy (R.T.) against cyclophosphamide, 5‐fluorouracil, and methotrexate (CMF) with and without postoperative irradiation therapy in decreasing the frequency of recurrence. No significant difference was observed between L‐PAM as compared with R.T. plus L‐PAM (P = 0.85). The difference between CMF and R.T. plus CMF was significant in support of CMF alone (P = 0.04). The frequency of recurrence between R.T. plus L‐PAM and R.T. plus CMF showed no difference. A comparison of the chemotherapy only regimens showed an advantage of CMF over L‐PAM (P = 0.05). Both the delay in starting chemotherapy and the significant decrease in percent optimal drug dosage during the first six cycles of therapy are factors that may influence the high frequency of relapse observed in the R.T. plus chemotherapy groups. Of the four treatments, CMF has the lowest frequency of recurrence.

Ultrasonic enhancement of nitrogen mustard cytotoxicity in mouse leukemia.

Mouse leukemia L1210 cells were exposed to continuous wave 2 MHz, 10 W/cm2 ultrasound for 10 minutes while suspended in nitrogen mustard solution in vitro. Mice subsequently inoculated with these cells had longer survival times than control animals that received cells exposed to the drug but not ultrasound. Without the drug, ultrasound did not alter survival time. Tracer studies revealed increased cellular accumulation of drug under the influence of ultrasound.

A new effective four-drug combination of CCNU (1-[2-chloroethyl]-3-cyclohexyl-1-nitrosourea) (NSC-79038), vinblastine, prednisone, and procarbazine for the treatment of advanced Hodgkin's disease

Five hundred and sixty‐six patients with either Stage III or IV Hodgkins disease were prospectively randomized to test whether CCNU and/or vinblastine are more effective than mechlorethamine and/or vincristine with procarbarine and prednisone. The combination of CCNU, vinblastine, procarbazine, and prednisone (CVPP) was shown to be a highly effective program with a complete response frequency of 69%. The use of CCNU as part of the induction program was also shown to be the most significant determinant of prolonged remissions (P = .025). Reduced vomiting and neurotoxicity, as well as the oral administration, were the chief advantages of the CVPP as compared with MOPP. These factors resulted in improved patient and physician compliance. The MVPP regimen was also shown to be a highly effective regimen with a complete response frequency of 73% in patients without prior exposure to chemotherapy. However, the induction regimens containing vinblastine were associated with a significantly higher frequency of fatal hematopoietic toxicities than the induction regimens containing vincristine (P = .05). This higher frequency was almost exclusively seen in the elderly or in patients previously treated with both chemotherapy and radiotherapy. At this time, the remission durations maintained by vinblastine with periodic reinforcement are longer when compared with vinblastine maintenance alone (P = .06), but there is no corresponding increase in survival.

Treatment of advanced non-Hodgkin's lymphoma with vincristine infusion

Twenty‐five patients with a variety of histologic types of advanced non‐Hodgkins lymphoma refractory to previous chemotherapy were entered into a trial of vincristine infusion. Patients received 5‐day courses of vincristine 0.25 mg/m2/day by continuous intravenous infusion after an initial 0.5 mg intravenous bolus injection. Courses were repeated every 3 weeks. Objective responses were observed in nine patients (36%), all of whom had previously received vincristine given by conventional bolus injection. A complete response occurred in a patient with diffuse mixed histiocytic lymphocytic lymphoma, and partial responses were observed in eight patients with the following histologic types: diffuse poorly differentiated lymphocytic (4); nodular poorly differentiated lymphocytic (2); diffuse mixed histiocytic lymphocytic (1); and diffuse histiocytic (1). Duration of response lasted from 1.2 to 16.2 months (mean, 4.4 months). The principal complication of therapy was mild‐to‐moderate neurotoxicity; this occurred in 12 patients (48%) who received a total of 54 courses of vincristine infusion. Hematologic toxicity was minimal and nausea/vomiting did not occur. Vincristine infusion may afford palliation for patients with advanced non‐Hodgkins lymphomas who have become refractory to standard chemotherapeutic regimens even if they have received prior vincristine by conventional bolus injection. These data suggest the possibility of enhancing the therapeutic efficacy of vincristine in the treatment of non‐Hodgkins lymphoma by use of an infusion technique.