Review of the Fialuridine (FIAU) Clinical Trials (1995)

This study was originally conceived in July 1989 to evaluate the effects of fiacitabine (FIAC), administered orally, for the treatment of cytomegalovirus (CMV) infection in human immunodeficiency virus-positive individuals. The study was developed by Oclassen Pharmaceuticals in conjunction with the AIDS Clinical Trials Group (ACTG) of the National Institute of Allergy and Infectious Diseases (NIAID). CMV is an important pathogen in immunocompromised patients such as those with HIV infection and solid-organ transplant recipients (Chazouilleres and Wright, in press). CMV is the most common infectious pathogen in HIV-positive patients, with the most frequent manifestation being chorioretinitis (occurring in 25-30 percent of patients) (Smith and Brennessel, 1994).

The prevalence of CMV disease following liver transplantation without antiviral prophylaxis is 15-41 percent (Chazouilleres and Wright, in press). Clinical manifestations include fever, malaise, arthralgias, pancytopenia, hepatitis, pneumonitis, gastrointestinal ulcerations, and chorioretinitis. Disseminated CMV disease is rare (occuring in less than 5 percent of those with CMV disease), but is ominous when it occurs since it is usually fatal (Chazouilleres and Wright, in press).

The drugs used to treat established CMV disease in HIV-positive patients include acyclovir, foscarnet, and ganciclovir. Ganciclovir has been used with the most success. Ganciclovir is a synthetic nucleoside analog of deoxyguanosine, which inhibits the replication of herpes virus both in vitro and in vivo (Matthews and Boehme, 1988). In CMV-infected cells, ganciclovir is phosphorylated and the metabolite inhibits viral DNA synthesis by competitive inhibition of viral DNA polymerase with direct incorporation into viral DNA and the resultant termination of viral DNA elongation (Erice, Jordan, Chace, Fletcher, Chinnock and Balfour, 1987). Foscarnet is another drug that has been proven to be effective against CMV (Smith and Brennessel, 1994). At the time that this protocol was developed, ganciclovir, which requires intravenous administration, had just become licensed; foscarnet, another intravenous drug with anti-CMV activity, was just beginning Phase II clinical trials. The need to identify an effective oral drug to treat CMV disease in this patient population was one of the most important scientific priorities of the ACTG.

FIAC was a promising drug. It demonstrated reasonable in vitro activity against CMV and excellent oral bioavailability. In July 1989, draft protocols designed to study the anti-CMV activity of FIAC were developed and circulated through several ACTG committees, including the CMV Pathogen Study Group, the Opportunistic Infection Core Committee, as well as the

NIAID Medical Branch and the Food and Drug Administration (FDA). The final protocol was established in October 1989.

The study was to be conducted at two clinical sites, the University of California at San Diego (Douglas D. Richman, principal investigator), and the University of Washington in Seattle (Lawrence Corey, principal investigator). Richman is an established investigator who has extensive experience in the study of antiviral agents used in the treatment of other herpes virus infections, such as herpes simplex virus types 1 and 2. He also is an established bench virologist with special interest in the pathogenesis of herpes virus infections. Richman was one of the lead investigators in the original studies that demonstrated the benefit of azidothymidine (zidovudine) in advanced HIV disease. On the basis of his experience investigating both herpes-related viruses and HIV, he was well suited to study the effects of a novel nucleoside agent such as FIAC in the treatment of HIV-associated CMV disease. Corey, likewise, is a clinician-scientist who has extensive experience studying herpes-related viruses in the laboratory as well as experience studying the use of nucleoside agents in the treatment of both herpes virus infections and HIV. Corey also had participated in the seminal studies of zidovudine for advanced HIV disease and was one of the lead investigators in studies which evaluated vidarabine and acyclovir for the treatment of herpes virus infections. Early in the HIV epidemic, Corey became established as one of the leading investigators in the treatment of AIDS-related diseases, and at the time that this protocol was being developed, he had assumed the primary leadership position within the ACTG as chair of its Executive Committee, which was responsible for oversight of the group's scientific effort.

FIAC study R89-001 (ACTG 122A) was a Phase I/II clinical trial designed to evaluate, simultaneously, the safety and relative anti-CMV activity of FIAC at doses of 0.6, 1.0, 1.7, 3.0, and 5.0 mg/kg/day given orally (roughly equivalent to the doses that MSKC found to be effective against VZV, assuming an average patient weight of 60 kg and a skin surface area of 1.3 m²). Patients with documented CMV viremia or viruria were to be entered successively into the treatment groups in an open-label fashion. A total of six patients had to have completed 14 days of therapy, with at least half of them having met protocol-defined tolerance criteria before patients could receive the next dosage level. Patients were of both sexes, but all were HIV infected and had CD4 counts of 500 cells/mm³ or less and CMV had been documented in either their blood or urine. All patients were followed for a total of 3 months, according to the protocol. The principal end points were the development of a negative CMV culture and tolerability of the drug.

On October 12, 1989, the study investigators, Oclassen Pharmaceuticals (which held the investigational new drug [IND] application), representatives from FDA, and other experienced virologists, including Carl Johnson and Richard Whitley, met to discuss the protocol. It was agreed to limit this study to a total of 4 weeks of therapy on the basis of the fact that data from animal toxicity tests of more than 90 days' duration were not available. The protocol received full institutional review board (IRB) approval at both sites plus a review and approval at the Veterans Administration Hospital at San Diego. Informed consent documents (ICDs) were also reviewed by these IRBs and were judged to be in compliance with national standards. The ICDs included a description of the potential known toxicities of FIAC, including bone marrow dysfunction, gastrointestinal disturbances, neurologic events (such as mental status changes and

seizures), and potential infertility. Appendix D contains copies of these ICDs and the ICDs from all of the FIAU trials.

One of the most striking features of this study was in the implementation of a novel mechanism of data and safety evaluation. On the basis of the experience of the two principal investigators in the development of other HIV-related therapeutic agents, such as zidovudine and zalcitabine, the investigators proposed a real-time data tracking system that employed transmission of laboratory and clinical event safety data to the study sponsor (Oclassen) via fax. Within a week after arrival at Oclassen, updated composite data summary sheets, compiled on a patient-by-patient basis, were faxed to all parties involved in the development of the protocol, including site investigators and coinvestigators, representatives from Oclassen, virology consultants (Johnson and Whitley), the NIAID Medical Officer (Judith Feinberg), and representatives from the FDA (Sandra Kweder and Bob Schnur). These faxed data summaries were followed by conference calls, held every 2 to 3 weeks. During these calls all new data were discussed, trends in the data were reviewed, and potential study changes and amendments were proposed and implemented. This data review process was used for both the R89 and the R90 studies, yielding a total of 92 distributed summaries (see Appendix E for an example) and approximately 30 conference calls. These procedures were not meant to supplant the usual adverse event reporting mechanisms or the 100 percent monitoring of all study charts performed by Oclassen. In addition to the fax data summaries and conference calls, the interim results and protocol amendments were reviewed and approved by the ACTG Viral Pathogen Study Group and the ACTG Opportunistic Infection Core Committee. Therefore, data safety review was conducted both internally among the study team and externally through an independent peer review process for both the R89 and R90 studies. Between October 1989 and March 1990, 10 patients were enrolled at the University of California, San Diego, study site. Several episodes of gastrointestinal side effects, predominantly nausea, as well as fatigue were noted in several patients receiving doses of 1 mg/kg. One patient had an asymptomatic elevation in the levels of creatine phosphokinase (CPK), a muscle enzyme, indicating potential muscle injury. In April 1990, patient 105 died of bacterial pneumonia 40 days after receiving the last dose of FIAC. Because of the difficulty in distinguishing clinical adverse experiences caused by drug versus those caused by the underlying disease, on April 16, 1990, the protocol team modified the study to allow only patients with CD4 counts of >200 cells/mm³ to be entered. Two patients, the first of whom developed fatigue and gastrointestinal symptoms, were enrolled at the University of Washington study site. Thereupon it was decided that, should additional patients develop significant symptoms, the study would be stopped. By mid-May, increasing reports of nausea and vomiting among study participants were noted, with no demonstrable anti-CMV effect, even at the 1.7-mg/kg dose level. At this point the study team recommended discontinuation of the development of FIAC as an anti-CMV drug and the study was stopped.

Three other patients died within 6 months of their last FIAC dose. Patient 103 died of AIDS-related Kaposi's sarcoma, and FIAC was not considered a causative factor. Patient 110 developed significant peripheral sensory-motor polyneuropathy, attributed to either HIV or CMV infection, shortly after finishing his last FIAC dose. His neurologic abnormalities were

complicated by multiple renal and electrolyte disturbances, and he died approximately 9 weeks after receiving his last dose of FIAC. Neither the investigators nor the 1993 FDA task force believed that this death was related to FIAC. Patient 107 was an individual with a history of hepatitis B virus infection and previous intravenous drug use. During his treatment with FIAC, he exhibited a minor elevation in transaminase levels and a slight elevation in CPK levels. Seven weeks after completing FIAC dosing, the patient was hospitalized with nausea, vomiting, and fatigue. Clinical chemistry evaluation revealed persistent elevations in AST/ALT levels (five to eight times baseline). He was taking zidovudine, methocarbamol, and amitriptyline, all of which were discontinued upon admission. The patient developed progressive hepatic failure, ascites, intrahepatic cholestasis, and, ultimately, hepatorenal syndrome and death 11 weeks after receiving his last dose of FIAC. Autopsy revealed hepatocellular necrosis, fatty metamorphosis, and cholestasis. The investigators concluded that this patient died of underlying disease or possibly drug-induced hepatitis because of amitriptyline or zidovudine. The 1993 FDA task force concluded that the death of this patient should, in retrospect, possibly be considered attributable to FIAC therapy.

Several features of this study are noteworthy. First, the study was carefully designed, with staggered patient enrollment and classic Phase I dose escalation. Second, the preclinical data and study design were reviewed prospectively and were approved both internally by the investigators and the sponsor and externally by several committees within ACTG and FDA. Most importantly, a real-time data assessment was performed via weekly faxed composite data summaries which were discussed by the investigators and sponsor during frequent conference calls. Although FDA was included on the list of fax recipients, the reviewers apparently found this level of uninterpreted raw data unhelpful (Feigel, 1994). The conference calls enabled the study team, however, to rapidly modify the protocol on the earliest indications of symptoms or possible antiviral activity. This process led to the determination of the maximum tolerated dose in humans after only 12 patients had been enrolled. Although this approach to real-time data assessment was relatively novel at the time that the trial was initiated, it has since been utilized in many Phase I studies and, ironically, is precisely in keeping with the changes to the Code of Federal Regulations recently proposed by the FDA, requiring semiannual reviews of composite safety data. In this study these data were reviewed approximately every 3 weeks.

With regard to patient deaths, three of the four patients who died within 6 months of the termination of FIAC were victims of HIV-related complications. The fourth patient, patient 107, died 11 weeks after receiving his last dose of FIAC. The 1993 FDA task force report classifies this case as one in which FIAC treatment should be considered a possible cause of death, but even in retrospect, although FIAC may have contributed to the observed liver failure, the confounding variables of CMV, hepatitis B and continuing treatment of HIV with zidovudine clearly preclude unequivocal classification of this death.