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

On the basis of the encouraging findings in the R90 dose-deescalation study, much enthusiasm was generated for instituting a study that evaluated more prolonged administration of fialuridine (FIAU) in patients with chronic HBV infection who were not coinfected with HIV. This study was conducted solely at the NIH Clinical Center, with Jay Hoofnagle as the principal investigator. Hoofnagle is a very experienced gastroenterologist and hepatologist who has developed a special interest in therapeutic approaches for viral hepatitis, especially hepatitis B and hepatitis C viruses. He is Director of the Division of Digestive Diseases and Nutrition at the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) at NIH. He was integrally involved in the early trials of vidarabine as a potential treatment of HBV infection and was the principal investigator on the major studies that established the therapeutic benefit of interferon in patients with chronic active HBV infection. He was one of the first investigators to describe the flare in transaminase levels associated with the successful clearance of HBV from the blood and has established a large group of patients with chronic active HBV whom he follows at the NIH Clinical Center.

In conjunction with Oclassen Pharmaceuticals, Hoofnagle and Stephen Straus developed a protocol entitled "A Four Week Course of FIAU for Chronic Hepatitis-B" that was submitted to the NIDDK IRB in August 1991. This study was initially designed to evaluate several doses of FIAU that would be administered over a prolonged period of time; however, the duration of therapy was ultimately limited to 28 days because there was only limited animal toxicity data to support longer-term courses of treatment. The rationale behind the desire for more prolonged courses of treatment was based on the understanding that HBV disease is a persistent infection associated with chronic viral replication. It was the opinion of both Hoofnagle and Straus that ultimate control of HBV disease would require some degree of prolonged chronic suppression for a durable effect. The standard treatment regimen for the only approved drug therapy, interferon, for example, calls for 16 weeks of daily subcutaneous injections. The promise of an orally bioavailable drug that was more potent than any previous anti-HBV therapy and that was associated with relatively few side effects at the lower doses generated much enthusiasm for this protocol. The protocol received approval, after compliance with a number of stipulations, by the NIDDK IRB and by the acting director of the Clinical Center. As part of their review, the IRB evaluated the informed consent document and found it to be in keeping with NIH standards. In the Risks and Discomforts section, fatigue, nausea, skin rash, bone marrow suppression, seizures, and "pains in the arms and legs due to irritation of the nerves or the muscles" were all described as potential side effects (Appendix D).

In February 1992 the protocol was submitted to FDA. During the time of the FDA review, information became available from the R90 study that suggested that lower dosing levels (e.g., 0.1 to 0.5 mg/kg/day) would be effective. As such, the protocol was amended to utilize four doses, 0.05, 0.1, 0.25, and 0.5 mg/kg, to be taken every day for 4 weeks. It was anticipated that six patients would be enrolled in each dosing cohort. This protocol was reviewed and approved by the NIH Clinical Center, and the first patient was dosed on April 18, 1992. By August 15, 1992, all 24 subjects had been enrolled in the protocol and all but 6 subjects had completed therapy. The length of clinical follow-up as stipulated by the protocol was 6 months. All patients ultimately completed therapy and followup. All patients returned for all study visits and were intensively monitored during and after therapy, including evaluation with a pretreatment liver biopsy (a posttherapy biopsy was considered but rejected as providing too little unique information to justify the known risks).

HBV DNA was suppressed 70 to 95 percent in the three groups receiving the highest dosage and to a lesser degree in the groups receiving the lowest dosage (0.05 mg/kg/day). Side effects during therapy were mild and transient. No patients developed side effects severe enough to modify the dose or interrupt therapy. In nine patients (patients 2A, 4A, 6A, 1B, 3B, 5B, 4D, 5D, 6D) HBV DNA levels decreased markedly, and two of these (patients 1B and 5B) eliminated HBV envelope antigen (HBeAg). Among the nine patients who became HBV DNA negative, flares in AST and ALT were noted in eight patients as the HBV DNA levels fell. Transaminase levels fell into the normal range in six of nine individuals and into the near normal range (45 to 61 units/liter) in the remaining three patients. With more than 1 year of follow-up, six of the nine patients have evidence of reactivated HBV disease with a return of HBV DNA; however, three patients have become HBeAg negative and have normal transaminase levels.

One patient (patient 4D) died within 6 months of receiving FIAU. This 59-year-old man with chronic HBV infection died on January 6, 1993, 2 months after laparoscopic cholecystectomy and 4 months after completion of his 28-day FIAU therapy. He had previously been treated in a 1989 interferon protocol at NIH (a pretreatment liver biopsy in 1989 revealed mild chronic active hepatitis with bridging fibrosis). He was subsequently enrolled in a prednisone/interferon trial at NIH, but he failed to have a sustained response to treatment.

In August 1992 he began FIAU as part of the R91 trial. A pretreatment liver biopsy in July 1992 showed chronic active hepatitis with fibrosis. His pre-treatment ALT level was 72 units/liter and his bilirubin level was 0.6 mg/dl. He received 28 days of FIAU until September 2, 1992 (total dose, 1,756 mg). During the last 2 weeks of treatment he developed mild nausea, fatigue, and myalgias. With treatment his HBV DNA levels fell from 52 to 1.9 pg/ml.

He was seen at NIH during a follow-up visit in September 1992 and was asymptomatic. However, in October 1992 he developed fatigue, nausea, and epigastric discomfort for which he was seen by his local physician. An upper gastrointestinal series and ultrasound were normal. A HIDA scan showed decreased gallbladder contraction after cholecystokinin. He was seen at NIH on October 25, 1992. His ALT level was two- to fourfold above baseline, he was HBV surface antigen (HBsAg) and HBeAg positive, but HBV DNA was low. Ultrasound and computed tomography scan again showed no evidence of gallstones and a heterogeneous liver. He was believed to be having a flare of hepatitis in response to FIAU. He complained of numbness in the feet and had clinical evidence of a neuropathy. Against the advice of the NIH

physicians, he underwent laparoscopic cholecystectomy in November 1992. No gallstones were present. An intraoperative liver biopsy showed chronic active hepatitis with fibrosis with a small amount of fat. Stains for HBV antigens were positive, but their levels were markedly reduced compared with those at the previous biopsy.

He was seen at NIH 1 week later, at which time he had developed ascites. A peripheral neuropathy was documented by nerve conduction studies. He continued to have ascites despite the use of diuretics and paracenteses. He was transferred from North Carolina to NIH on December 20, 1992, 4 months after completion of his FIAU treatment, at which time he had massive ascites and intravascular volume depletion. Hepatic function was preserved (bilirubin, 0.8 mg/dl; albumin, 3.5 g/dl; prothrombin time, normal). Refractory ascites was ascribed to complications of the cholecystectomy. He had also developed lactic acidosis of unknown etiology and had biochemical evidence of pancreatitis (normal amylase levels but high lipase levels) but no abdominal pain. Despite normal serum bilirubin and near normal liver enzyme levels, he was encephalopathic and was believed to be in hepatic failure. He was therefore transferred to Pittsburgh for liver transplant evaluation. He died of multisystem organ failure and lactic acidosis prior to transplantation. An autopsy performed at NIH revealed inactive cirrhosis of the liver with marked microvesicular steatosis, but rare hepatocyte staining for HBV antigens. The pancreas was autolyzed but had histologic evidence of pancreatitis.

The etiology of this patient's syndrome was puzzling, but was not thought to be related to FIAU at the time since the patient was no longer on the drug and hepatic decompensation appeared to be temporally related to cholecystectomy. Nevertheless, a description of this death was included in the Clinical Investigator's Brochure from Eli Lilly dated February 1993 which was central to the preparation of the ill-fated 6-month protocol (PPPC)(Lilly Research Laboratories, 1993). The death was ascribed to complications of surgery in that document, although a simple "cause unknown" would probably have been a better reflection of reality. It was also discussed by medical staff fellows with every patient who participated in the PPPC trial.

Several other patients in the R91 study developed evidence of hepatic or pancreatic damage within 6 months after receiving their last dose of FIAU. Patient 006B was a 59-year-old, HBV-infected male who received 0.25 mg of FIAU per kg/day for 28 days starting in late May 1992. Approximately 4 months after the discontinuation of FIAU therapy, the patient was hospitalized for severe abdominal pain. The pain was not associated with an increased ALT level or any abnormalities of amylase, lipase, or acidosis, and no diagnosis was established. This patient still experiences bouts of similar pain without explanation. The hospitalization was reported to the study sponsor in a routine fashion, but the investigators did not feel that this abdominal pain was related to the FIAU therapy.

Two additional patients experienced significant health problems in the months following their participation in the R91 trial. Patient 1C is a 56-year-old HBV-infected male who received 0.05 mg of FIAU per kg/day in June 1992. He was admitted to the hospital in October 1992 with right upper quadrant pain, at which time a cholecystectomy was performed. A liver biopsy obtained perioperatively showed moderately active micronodular cirrhosis; a stone was found in the cystic duct. In the opinion of the investigators, the hospitalization appeared to be due to a preexisting problem (cholelithiasis that was known before his FIAU therapy) and unrelated to FIAU.

Patient 1A is a 52-year-old man who received 28 days of FIAU therapy beginning on April 13, 1992, but he developed sensations of numbness and tingling in late September 1992 (4 months after receiving his last FIAU). Because of progressive pain, he underwent nerve conduction studies at NIH in January 1993. These were compatible with a peripheral neuropathy. Symptoms were similar to those experienced 5 years previously when he had been consuming significant amounts of alcohol. Because of this prior history and because the neuropathy was believed to be atypical for a nucleoside analog in that it began 4 months after FIAU treatment was stopped and persisted (Ara-A induced neuropathies, for example, are usually transient), the investigators could not draw a clear relationship between drug and toxicity. In retrospect it is likely that this neuropathy was FIAU related. At the time this was considered to be a possibility by the investigators, and the appropriate precautions were taken to monitor subsequent patients. Because of this patient and the unequivocal peripheral neuropathy of patient 4D (above), peripheral neuropathy was included as a potential complication of FIAU in the PPPC trial and all patients were monitored with nerve conduction studies.

This study clearly established the clinical utility of FIAU in the treatment of chronic HBV infection. All patients responded with reductions in circulating HBV DNA levels, and many in the groups receiving the three highest doses became HBV DNA negative. Unfortunately, this antiviral effect was not maintained for an extended period after the cessation of drug therapy. This lack of permanence and the relative absence of dose-modifying toxicities while the patients were being treated strongly suggested the need for a longer duration of FIAU treatment, which was the rationale for the 6-month PPPC trial.

As with the previous R89 and R90 studies, several patients developed clinical problems after finishing their courses of treatment. In the cases of patients 6B and 1C, it is most likely that their clinical problems were due to preexisting disorders and were not related to FIAU.

Patient 4D, on the other hand, presented with a more complicated clinical picture. In retrospect it is apparent that this patient died of what would later be termed the FIAU syndrome with systemic lactic acidosis and hepatic steatosis. On the basis of the unchanged liver biopsy findings at the time of the patient's cholecystectomy, the investigators concluded that his subsequent liver failure and death were due to complications of the cholecystectomy, possibly related to the anesthetic agent, rather than to delayed toxicity of FIAU. Since the syndrome of FIAU toxicity was undescribed at the time, the investigators cannot be strongly criticized for coming to this conclusion. Indeed, it is obvious that the investigators took this death very seriously. The progress and course of this patient were fully described to the sponsor via telephone. The death was reported to the NIDDK IRB, was discussed at NIH senior staff meetings, reviewed by pathologists at the Armed Forces Institute of Pathology, and described in the subsequent PPPC protocol. After discussion with the NIDDK IRB, the death was not described in the consent form given to participants in the PPPC trial, but it was nevertheless discussed with each patient before they signed that form. Furthermore, this patient's clinical course, along with a detailed review of four other known deaths of former FIAU- or FIAC-

treated patients, were reviewed at a pre-PPPC protocol meeting with FDA in April 1993. FDA approved the initiation of the PPPC trial after those discussions.

The attention paid to this patient's death by the investigators and their acute awareness of the peculiar nature of his presentation were no doubt responsible for their increased vigilance throughout the conduct of the ensuing PPPC study and the rapid termination of that trial when a patient presented with a similar syndrome on June 26, 1993.