Further considerations on the evaluation of potential reduced-risk tobacco products. Part II: Re-assessment of a heuristic using the CPS-II database

In a previous analysis (see Part I) we proposed a heuristic for assessing the efficacy of potential reduced-risk tobacco products (PRRPs) on lung cancer (LC) rates, using smoking cessation data published in a report from the Iowa Women’s Health Study (IWHS) as a basis for sample size estimates. In this study, an additional analysis was performed using cessation data from the much larger Cancer Prevention Study II (CPS-II), which also provides data on different durations of cessation. Statistical methods were used to assess whether smokers switching to a PRRP would reduce their risk of LC. Furthermore, non-inferiority tests compared the LC risk in switchers to that in smokers who had quit smoking. The present work shows that similar sample size estimates were obtained whether the analysis was based on the IWHS or the CPS-II data sets, suggesting that the heuristic may be generally applicable to prospective real-life studies to evaluate PRRPs. Non-inferiority testing of switchers compared with quitters required approximately 10-fold more subjects than did superiority testing of switchers compared with smokers. Altogether, these estimates indicate that it is feasible, in terms of study duration and sample size, to clinically assess the LC risk-reducing potential of a PRRP.