Cancer prevention trials: Assume nothing, expect everything.

ED05-02

The implementation of all clinical trials is challenging. However, because the population targeted in prevention trials is healthy volunteers, the size of prevention trials is so large, and because they are conducted over an extended time frame, these trials pose additional complexities. Trials of this magnitude involve the cooperation of a number of agencies, clinical centers and outside vendors. Events external to the trial itself, including changes in the medical community or external publications using the study agents, can impact the study. Despite careful planning, there are bound to be situations which could not have been anticipated. The success of running these trials depends on the response of the study team to these circumstances.

On the Prostate Cancer Prevention Trial (PCPT) the use of a study agent finasteride affected the endpoint resulting in challenges from the original design phase, through implementation as well as at analysis and during the interpretation of the study results. At the inception of the trial in 1992, the use of prostate specific antigen (PSA) was just beginning to be widely used. Finasteride was known to effect PSA levels such that the level for a man on finasteride was approximately halved. To compensate for this differential PSA level and to maintain the study blind, it was decided that the PSA levels would be analyzed centrally and that the PSA results would be reported as "Elevated" (> 4.0 ng/ml) or "Not Elevated" (= 4.0 ng/ml). Early on in the trial it became clear that this was reporting mechanism was not going to be accepted by the study community. Because the reporting of the PSA results had the potential of affecting the study blind, the Data and Safety Monitoring Committee (DSMC) became involved in discussions regarding how to best handle the situation. A recommendation by the DSMC was made to the Steering Committee and approved. The change in reporting would be such that if the value was = 4.0 ng/ml it was reported as "Not Elevated" and if > 4.0 ng/ml, the value, adjusted for finasteride use, was reported. This satisfied the study sites for two years. However, as the use of PSA velocity became popular even if the adjusted PSA was = 4.0 ng/ml, it then became clear that study sites were interested in the actual values. A second change in reporting was recommended by the DSMC to the Steering Committee and approved. Study sites could get a PSA history for a participant but only by request. These requests were monitored by Statistical Center personnel. The combined response by the DSMC, Steering Committee and study sites ensured that the safety of the participants was not compromised while making sure that all efforts were made to maintain the study blind.

The Selenium and Vitamin E Cancer Prevention Trial (SELECT) is a prostate cancer prevention study which uses two readily available agents. In addition, these two agents are included in all brands of commercially available multivitamins. In order to keep the participants from taking excess vitamin E and selenium in their daily multivitamin, it was decided that the study would provide a specially made multivitamin without either vitamin E or selenium. A manufacturer was found who was willing to supply the entire study population (expected to be 32,000 recruited over five years) for the entire length of the study (projected median follow-up of 9 years) for free The formulation was agreed upon by the manufacturer as well as the trial designers. An initial shipment of the multivitamin specially formulated for SELECT was delivered to our Pharmacy Coordinating Center. Within four months of the first randomization, it became clear that the manufacturer had over-estimated its ability to provide the promised multivitamin. As most of the initial shipment had been shipped to the sites based on initial accrual estimated and the promised second shipment didn't arrive as initially promised, it became necessary to reallocate the available multivitamin among the study sites. This required study sites to reship the multivitamin to the Statistical Center which then resent the multivitamin to sites in need. While the manufacturer began to get paid for their product, it became clear to the trial leadership that the current arrangement with the multivitamin manufacturer would not work. A Request for Application (RFA) was distributed to multivitamin manufacturers. This RFA also provided us with an opportunity to re-formulate the multivitamin as it became clear that the vitamin D content was not considered to be adequate and was a problem with some of the study sites' IRBs. Eventually an alternate manufacturer who was able to provide a superior product at a price that was more than "free" but less than that of the initial manufacturer was found.

These examples, as well as other examples from the PCPT, SELECT and other cancer prevention trials will be used to illustrate the dynamic nature of long term prevention trials.