Filter analysis for evaluating the mouth level exposure (MLE) of smokers is non-invasive, allowing ordinary smoking behaviour with simple sample collection. We investigated 1) estimated MLEs to 36 smoke constituents using filter analysis, 2) MLEs of Japanese smokers smoking cigarettes with different tar content, and 3) the relationship between their MLEs and index values obtained by machine smoking.

A low tar cigarette (LTC, 6 mg ISO tar) and an ultra-low tar cigarette (ULTC, 1 mg ISO tar) were smoked using a smoking machine (18 regimes). The nicotine content in the filter butts and 36 constituents such as tar, nicotine, and CO were measured. Except for formaldehyde, a strong correlation was found between the filter nicotine content and each of the 35 constituents in both cigarettes (R² = 0.78 - 0.99). This result shows that the MLEs to 35 constituents could be estimated from quantification of the nicotine content in filters.

Japanese male smokers (LTC smokers: 105, ULTC smokers: 105) aged 21-49 years were recruited. Subjects were permitted to smoke their regular use cigarettes ad libitum for 7 hours in the examination site. After collection, the nicotine content in each filter was measured to estimate the MLEs. The estimated MLEs to 35 constituents in ULTC smokers were significantly lower than in LTC smokers.

The relationships between the estimated MLE to each constituent from LTC and ULTC smokers and five index values obtained by machine smoking (ISO and Canadian Intense) – “each constituent”, “tar”, “nicotine”, “each constituent/tar”, and “each constituent/nicotine” – were evaluated. The results showed that relationships varied among the indexes.