Our recently concluded Phase-IIB trial of Bowman-Birk Inhibitor Concentrate (BBIC) in patients with oral leukoplakia was negative, despite a foundation of earlier work suggesting BBIC has chemoprevention potential. As presented at the 2010 Frontiers meetings, there was no statistically significant difference in relative percent change in total lesion area between the BBIC arm and the placebo arm after six months of therapy. This result is unexpected, given both our earlier demonstration of a dose-response relationship in a Phase-IIA trial, and the interim analysis of the Phase-IIB trial. The aim of this presentation is to illuminate two factors possibly contributing to the negative final results: 1) an apparently active placebo, and 2) a change across time in the potency of the BBIC dispensed. We present here admittedly post-hoc, but thought-provoking, analyses to support these notions.

BBIC has protease-inhibitor activity, but any such activity in the placebo was unanticipated. The interim analysis of the Phase-IIB trial showed the 95-percent interval estimate of relative percent change in total lesion area straddled zero in the placebo arm (95%CI: −39.0 to +4.5, N=30). In contrast, the corresponding interval for the BBIC arm was entirely below zero (95%CI: −49.3 to −10.9, N=25). Thus the interim results seemed to predict a positive final result. The final analysis, however, showed the mean relative percent change in total lesion area was significantly less than zero in each of the two study arms (placebo mean: −17.1, N=46; BBIC mean: −20.6, N=43; both sign-rank tests p < 0.05). Reassessment of the placebo showed <10 chymotrypsin inhibitor units/gram and therefore the changes in total lesion area among placebo subjects were likely to be due to other mechanisms of action in the placebo.

The juxtaposition of the interim and final results led us to examine data in the BBIC arm by randomization date. Relative percent reduction in total lesion area was greater for subjects enrolled earlier, relative to those enrolled later (Spearman r = 0.31, p < 0.05, N=43), consistent with a decline in agent potency. (There was no such trend in the placebo group: Spearman r=0.22, p > 0.44, N=46). Of necessity, different batches of BBIC were used in the trial. It is known that the potency of BBIC can be greatly reduced by maintenance at refrigerated or frozen temperatures, and that some companies have stored BBIC under such conditions. We hypothesize that the BBIC batches used later in this trial had reduced potency, possibly due to storage conditions that compromised its ability to serve as a cancer chemoprevention agent by reducing bioavailability of the drug. Our data provide evidence that the later batches of BBIC had compromised chemoprevention activity. Median relative percent change in total lesion area was −32% (N=36) for the earlier batches but +6.5% (N=7) for the batches dispensed later in the trial. There was a similar pattern in relative percent change in buccal-cell neu protein (older batches median:-16.6, N=14; newer batches median: +24.6, N=7), relative percent change in protease activity (older batches median: +1.6, N=14; newer batches median +136.5, N=7), and relative percent change in serum neu (older batches median: −4.1, N=13; newer batches median −2.6, N=6). While many factors contribute to trial outcome, these data suggest our Phase-IIB trial was affected by the reduced potency of BBIC over time. Further, the apparent, and not understood, activity of the placebo raised the bar for declaring BBIC a success in this context.

Citation Information: Cancer Prev Res 2011;4(10 Suppl):B82.