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

Anyone apprised of the disastrous events that occurred in the PPPC trial at NIH can only have been shocked and saddened by the suffering of the seven patients who developed sudden and severe FIAU toxicity (lactic acidosis, liver failure, pancreatitis). The Institute of Medicine committee understands the grief that must have been felt at the time by the families and close friends of the 5 individuals who died, all volunteers for a clinical trial for which hopes were high for clinical benefit; and the committee recognizes that their families undoubtedly continue to carry a heavy emotional burden. The recommendations of this IOM committee, as the earlier recommendations of the NIH subcommittee and the FDA, have the goal of reducing to a minimum the possible occurrence of such a calamity in any future clinical drug trials, but their implementation cannot render the process absolutely risk free. At the same time we believe that the recommendations are not so stringent or onerous as to restrict the benefits to the ill that can accrue from successful new drug development.

The deaths and morbid events related to the PPPC trial of FIAU have, as with other tragedies, resulted in private and public soul searching as to their cause and as to whether they were preventable. There have been efforts to affix blame. There have been calls for new safeguards and procedures to protect against future tragedies of the kind described here without full consideration as to whether those recommendations are adequately supported by an analysis of the costs and benefits of such safeguards. Although the history of drug regulation has been driven, in part, by disasters that have prompted new legislation (Hutt, 1994), that tendency should be resisted in this case because there is no evidence that anything could have been done to protect against the previously unknown and unanticipated late toxicity of FIAU.

The United States is a culture in which an event perceived by the public as evil, tragic, or unjust evokes often the response, "There ought to be a law!" (Toulmin, 1981). It may be useful to consider such "evils" or ''tragedies" as falling into two classes. The first is the kind of evil that can be remedied or prevented by human action. In such cases it is appropriate (and sometimes morally obligatory) to consider the development of public policy designed to remedy or prevent the evil in question. A good case in point was the finding in September 1937 that more than 100 people died of diethylene glycol poisoning following the use of sulfanilamide elixir, a preparation in which the antibacterial drug sulfanilamide was dissolved in diethylene glycol for pediatric use. This solvent had been marketed without any prior safety testing (Hutt, 1994). In response to this, Congress added to the Federal Food, Drug, and Cosmetic Act of 1938 a requirement that new drugs be tested for safety in humans before being licensed for commercial distribution.

The second category of evil consists of tragic or unfortunate events that cannot be remedied or prevented by any reasonable human behavior. The appropriate response to such events is to acknowledge their existence and to view them with fear and trembling, but to refrain from a regulatory reaction until an action that can prevent their future occurrence is clear. A good case in point is the lethal reaction to FIAU. At the time of this writing the IOM committee is aware of no test that could be done to anticipate such lethal reactions. Among the actions we recommend is the conduct of careful studies aimed at developing a test or series of tests that would identify other agents that could produce such reactions. If and when such a test becomes available, it should be considered a part of routine preclinical testing of new drugs.

The public's frequent tendency to rely on a regulatory response to tragedies of this type may be grounded in some incorrect assumptions. Some individuals appear to assume that all calamities are preventable. Therefore, when they occur it is further assumed that somebody must have been careless, negligent, or exploitative. Favorite targets for such accusations are the industrial sponsors, the investigators, and the regulatory agency. Such behavior by the sponsors is usually attributed to their presumed willingness to override all other values in the single-minded pursuit of corporate profit. Investigators are similarly accused of negligent behavior owing to their presumed overly-zealous pursuit of financial or academic rewards such as tenure and prestige (e.g., Twentieth Century Fund Task Force, 1984). Regulatory agencies are portrayed as inept or out of touch with the needs of the people, and individuals within the agencies are accused or suspected of conflicts of interest.

The committee wishes to call attention to two dangerous consequences of these assumptions. The tendency to "close the book" by indicating blame has the effect of reinforcing the erroneous public perception that new treatments can be developed and tested free of risks if only enough care is taken. As a result, once society is satisfied that it has identified the individual or group responsible for the tragic event and removed this danger from the drug development system, we can all feel free to return to business as usual.

A second danger is even more insidious. The failure of the public to appreciate the risks of research in the field of drug development compounded with a strong tendency to affix blame on those who carry out the trials when something goes wrong creates a risk for the future of clinical research. Young medical scientists cannot fail to notice that this situation presents grave risks to their personal reputations and careers. Skilled clinical investigators in the field of drug development are a valuable national resource who can reduce the burden of pain, illness, and premature death.

The committee has found nothing in its review to suggest that the tragedy was preventable given the heretofore unobserved and unanticipated late complication of the treatment. Furthermore, we find nothing to suggest that investigators were negligent in the conduct of the trials under review or that they were calloused or insensitive to the needs and conditions of the patients whom they studied.

The fact that we cannot at this time recommend a regulatory action that would absolutely prevent a recurrence of a tragic event of the sort caused by FIAU should not be a source of intimidation with respect to new drug development. Examination of the past record of experience with Phase I testing of new drugs shows that, in fields other than oncology,

lethal or disabling adverse events are exceedingly rare. For example, Zarafonetis, Riley, Willis, Power, Werbelow, Farhat, Beckwith and Marks (1978) found that in Phase I drug testing in prisoners a "clinically significant medical event" occurred once every 26.3 years of individual subject exposure. In 805 protocols involving 29,162 prisoner subjects over 614,534 days, there were 58 adverse drug reactions, of which none produced death or permanent disability. The only subject who died did so while receiving a placebo.

Cardon, Dommel and Trumble (1976) reported the results of their large scale survey of investigators designed to determine the incidence of injuries to research subjects. They found that the risk of either disability (temporary or permanent) or of fatality was substantially less than the risk of similar unfortunate outcomes in similar medical settings involving no research. Further examples illustrating the safety of being a research subject can be found in Levine (1986) and Williams(1990).

To summarize, on review of the fialuridine tragedy, the IOM committee finds no evidence of negligence or carelessness on the part of the investigators or sponsors. The committee is not aware of any preclinical tests that could have been done to anticipate this unfortunate outcome. This should not be a source of great alarm to the public, since there has been an excellent record of safety in early phases of research in the field of drug development in the United States. Against this background, the FIAU tragedy can be put in proper perspective as a painful but distinctly rare occurrence.

As part of this review of the fialuridine studies, the IOM committee will make some recommendations on the future conduct of early-phase drug trials. These are recommendations for the "fine- tuning" of rather than for making major changes in the drug development process. Although the IOM committee's recommendations are focused on Phase I and Phase II trials, they do not stem from perceived deficiencies in the conduct of the investigators or sponsors in the FIAU studies (on the basis of the current knowledge); as already noted, no deficiencies were identified. The IOM committee believes that implementation of its recommendations may reduce the already very low probability of occurrence of toxicity of the sort that occurred in the FIAU trials, particularly when studying drugs that are similar to FIAU in structure or action. In many instances the effect will be to insure the continued and consistent use of procedures and practices employed in the FIAU trials reviewed here.

1. Proceed cautiously in revision of the current drug development system. The current system has evolved checks and balances that have benefited new drug development, and no single component can undergo a major revision without endangering the system as a whole. The committee is doubtful that any of the changes reviewed above and/or proposed below, had they been in effect before the initiation of the FIAC and FIAU trials reviewed here, would have substantially altered the tragic outcome.

2. All clinical researchers engaged in trials should be exposed to explicit training not only on the design and conduct of clinical trials and their ethical obligations to patients but also on their legal and regulatory obligations to both the sponsor and the FDA.

3. We urge the establishment of a system of no-fault compensation for research injury by government, sponsors, or some combination of both.

4. Some form of independent safety monitoring would be a valuable component of any clinical trial in which patients are treated for extended periods, but they are especially important for all double-blind trials and in any trial in which there is reason to anticipate that evidence of adverse reactions could be confused with evidence of disease progression or therapeutic response. For other types of trials the sponsor should bear the burden of demonstrating that a monitor is unnecessary.

5. We support in principle the desirability of controls in Phase II studies, even while recognizing that statistical power will generally be inadequate to detect all but the most common of adverse drug effects, and among those, only those that rarely occur in untreated subjects. Nevertheless, particularly in trials involving extended treatment of patients, the use of some concurrent comparison group should help focus attention on the importance of differentiating drug effects from the underlying disease(s).

6. Research into the development of a database from which to construct historical control groups should be supported. Such control groups may be needed for comparisons when suitably matched concurrent control groups are not feasible. The extensive data submitted to FDA through the IND process are a potentially valuable resource for custom matching patients in new drug trials with controls from previous trials, matching not only for entry criteria but also for disease extent and severity, concomitant medications and other confounding variables.

7. Concerted efforts should be made to include in all clinical trial protocols explicit prospective criteria to help distinguish between adverse events related to drug treatment and changes in the underlying disease, for better or worse, whether or not controls are employed.

8. Clinical protocols should also have a section explicitly addressing the determination of the followup period, based on preclinical data and clinical data from other drugs thought to be similar in structure and action. Drugs suspected of modifying nucleic acids or associated macromolecules demand a minimum of 6 months of followup.

9. At the outset of extended Phase II trials, consideration should be given as to whether there is sufficient evidence of safety to justify simultaneous enrollment of a substantial group of patients, or whether the patients' disease is so serious that access to the drug seems warranted.

10. Data should ideally be analyzed (by the investigators or independent safety monitor in Phase I and Phase II trials and by data safety monitoring committees or data coordinating

centers in multicenter Phase III studies) on a continuing real-time basis rather than only after all case report forms are complete for all patients. This will ensure that the fewest possible patients are exposed to possible hazards and that rapid intervention will prevent or limit injury to individual patients.

11. We concur with the suggestion of the FDA Task Force that some form of cumulative adverse event reporting should be provided by the sponsor in a form that includes not only those events previously reported as serious, unexpected, and drug related but also any events judged to have met only the first or the first two of those conditions, along with the sponsor's explanation of the event. A careful analysis of all available information rather than a worst-case assumption should then determine further actions.

12. The committee believes that requiring a cumulative and all-encompassing report of the sort referred to in the previous recommendation every 6 months will prove to be a substantial impediment to development of drugs to combat life-threatening diseases (e.g., cancer) in which adverse events are frequent because of the often progressive nature of the underlying disease. Some judgment will always be necessary, by the investigators in deciding the most likely cause of adverse events, and by the FDA in deciding for which drugs and at what intervals a cumulative safety summary is necessary. Ideally, the investigators and FDA would work together in making both of these determinations.

13. Given the intense publicity surrounding the issue of FDA warning letters, regard for due process would seem to demand some modifications in procedure. A fairer process, for example, might be for FDA to refrain from public release of the letters until receipt of the mandated reply from the addressees. Some sort of appeal process, independent of the FDA Office of Compliance, and including scientific peers expert in the area, should also be available in the event of intractable disagreement, and when retrospective evaluation by the Office of Compliance is at odds with the prior evaluation of the responsible review division, that fact should at least be acknowledged in the letter.

14. We urge continued support for research into the mechanism of FIAU toxicity, including development of animal models and other test procedures that could be used to screen drugs with potential to alter DNA before performing clinical trials.

15. Preclinical testing in animals is especially important in the case of FIAU-like drugs with a potential for long-term effects on nuclear or mitochondrial DNA. Toxicology studies in at least two different species, using the route of administration intended for use in patients and a duration of treatment at least as long as that intended for clinical trials, and extended follow-up on at least a subsample of animals should all be key considerations in preclinical testing. Whenever feasible and reasonable, the preclinical assessment should include testing in an animal model of the disease being targeted.