Having a tool that can predict the Hoffmann analytes has become necessary, not only because of regulations, but also as a reference for the chemical composition of smoke for the Potential Reduced Exposure Products (PREPs).

The work presented in this paper is a systematic study on the role of cigarette paper main parameters, (porosity and citrate), and of filter ventilation on all of the Hoffmann analytes.

To conduct this systematic study, we have used a face centred central composite design. This type of matrix allows a global analysis on the parameters studied, taking into account possible relationship of the second order, and possible interactions between the parameters studied.

For the 3 levels of variations requested in these type of study we have used:

• Cigarette paper porosity: 20 ; 50 ; 80 Coresta
• Amount of Na/K citrate: 0.6 ; 1.3 ; 2.0 %
• Levels of filter ventilation: 0 ; 20 ; 40 %.

As the influence of the level of citrate in cigarette paper on CO and CO/Tar could be more complicated than a second order relationship, we have added at the centre point (50 Coresta, 20% ventilation), 2 levels of citrate to modelise citrate effect from 0% to 2.5%.

The tobacco blend used in this study is a standard American blend.

Cigarettes were smoked for all Hoffmann analytes, including tar, nicotine and carbon monoxide.

Data analysis from this study led to a number of conclusions, which will be presented.

These 2 studies give us some predictive models for the main analytes of Hoffmann's list in function of paper porosity, amount of citrate and filter filtration.

• Like for tar and CO, filter ventilation is a major tool to reduce most of the Hoffmann analytes; the porosity of cigarette paper, to a lesser extent, also plays a significant role.
• Concerning the prediction rules, we will classify the different analytes in groups : if most of them can be simply linked with either tar, nicotine or CO, some others show significantly different trends which will be shown. Tentative explanation will be given.