Formation of mainstream cigarette smoke constituents prioritized by the World Health Organization – Yield patterns observed in market surveys, clustering and inverse correlations

• Published in: Food and chemical toxicology Vol. 55; pp. 329 - 347
• Main Authors: Piadé, J.-J., Wajrock, S., Jaccard, G., Janeke, G.
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01.05.2013; Elsevier
• Subjects: acrolein; benzo(a)pyrene; Biological and medical sciences; chemistry; Cluster Analysis; Data Collection; formaldehyde; Humans; Machine-smoking regimen; Market map; markets; Medical sciences; Nicotiana; nitrogen oxides; Pattern recognition; Precursors; principal component analysis; public health; Smoke - analysis; Smoke chemistry; Smoking; smoking (habit); surveys; Tobacco, tobacco smoking; Toxicology; WHO mandated toxicants; World Health Organization; Virginia; USA, Virginia; Market map; Machine-smoking regimen; WHO mandated toxicants; Pattern recognition; Smoke chemistry; Precursors; Cartography; Tobacco smoke; Tobacco smoking; Precursor; Mainstream smoke; Survey; Maps; Addiction; International organization; Poison; Yield; WHO
• ISSN: 0278-6915; 1873-6351
• DOI: 10.1016/j.fct.2013.01.016

► Smoke constituents yields of Virginia and American blend brands are quite different. ► Such brands are well separated as clusters in a PCA using smoke yields as variables. ► Negative correlations of smoke yields are revealed by the PCA principal components. ► Yield pattern differences are explained on the basis of smoke formation processes. ► Differences in blend type are linked to observed negative correlations and clusters. The WHO TobReg proposed mandating ceilings on selected smoke constituents determined from the market-specific median of nicotine-normalized yield distributions. Data validating this regulatory concept were obtained from essentially single-blend surveys. This process is strongly impacted by inverse correlations among yields. In the present study, 18 priority WHO smoke constituent yields (nicotine-normalized) were determined (using two smoking regimens) from 262 commercial brands including American, Virginia and local blends from 13 countries. Principal Component Analysis was used to identify yields patterns, clustering of blend types and the inverse correlations causing these clusters. Three principal components explain about 75% of total data variability. PC1 was sensitive to the relative levels of gas- and particle-phase compounds. PC2 and PC3 cluster American- and Virginia-blends, revealing inverse correlations: Nitrogen oxides and amino- or nitroso-aromatic compounds inversely correlate to either formaldehyde and acrolein, or benzo(a)pyrene and di-hydroxybenzenes. These results can be explained by reviewing the processes determining each components smoke delivery. Regulatory initiatives simultaneously targeting selected smoke constituents in markets with mixed blend styles will be strongly impacted by the inverse correlations described. It is difficult to predict the ultimate impact of such regulations on public health, considering the complex chemistry of cigarette smoke formation.