The relationship between cigarette blend sugar and acetaldehyde formed in its smoke is a matter of current regulatory interest. Some tobacco scientists have used data by Phillpotts ^[1] to illustrate that cigarette smoke yields of acetaldehyde are independent of inherent or added sugar levels in the parent tobacco blends. Some tobacco control advocates (O'Connor and Hurley^[2]) have recently interpreted the data differently. In this work we re-evaluated the Phillpotts data and also similarly evaluated our own chemistry data across cigarettes representative of our full portfolio. Using the Phillpotts's data, O'Connor and Hurley recently applied a multivariate analysis to determine the relationship between acetaldehyde in smoke and sugar in tobacco taking into account the NFDPM yields. Indeed, a multivariate analysis enables the partial effect of different factors, especially product design, to be taken into account. With this analysis normalized for NFDPM, the authors concluded that the sugar-acetaldehyde relationship had a coefficient of correlation of 34%. However, the authors decided to use only the data which come from filter cigarettes whereas it would be more relevant to consider the full dataset and to add in the model a qualitative factor with two modalities (FP: filter and plain cigarette). Moreover, the difference between cigarettes sold in different countries has not been taken into account in their model although Phillpotts mentioned the specific behaviours of Italian and French brands. Therefore, the multivariate analysis used by O'Connor is incomplete. In order to avoid any misleading conclusions, we decided to carry out a multivariate analysis (General Linear Model (GLM)) taking into account all the factors (Sugar, NFDPM, FP and Country). Contrary to the O'Connor conclusion, our results have shown the sugar content does not have a significant impact on aldehyde yield.

In a second step, we have examined some data sets obtained in our laboratory between 2001 and 2010 on 99 commercial brands from the EU market: 12 from Eastern Europe (Poland, Hungary, Ukraine), 34 from France, 14 from Germany, 8 from Spain and 31 from the UK. In our studies, a GLM analysis with the sugar factor nested in the country factor has shown no effect of sugar content on acetaldehyde yields whatever the country. Moreover, an analysis taking into account the blend style has also highlighted no relationship.

[1] D. F. Phillpotts, D. Spincer, D. T. Westcott. The effect of the natural sugar content of tobacco upon the acetaldehyde concentration found in cigarette smoke. Beitraege zur Tabakforschung, 8 (1975) 7
[2] R. J. O'Connor, P. J. Hurley. Existing technologies to reduce specific toxicant emissions in cigarette smoke. Tobacco Control, 18 (2008) 139.