We have developed and used part-filter analysis to estimate human mouth-level exposure (MLE), and variants of this method are being reported by researchers outside of the tobacco industry[2,3]. Biomarkers of exposure (BoE) are also widely acknowledged to be an effective way of measuring human exposure to smoke toxicants.
Each approach has its advantages and applications. Filter analysis is minimally invasive and is well suited to estimating MLE in medium - to large-scale studies. Studies involving biomarkers are more invasive but can provide data on the mechanisms and, potentially, the biological significance of exposure reduction.
To provide confidence in the data derived from these two methods, we deisgned, conducted and published the findings of a study in which estimates of smoke toxicant exposure were obtained using both methods, and the relationship between these estimates assessed by way of a correlation[4,5].
This study was approved by an independent ethics committee and conducted in a clinic in Germany. Two hundred healthy volunteers were recruited comprising 50 smokers of each of 10mg, 4mg and 1mg ISO tar yield commercially available cigarettes plus 50 non-smokers. This population was expected to provide a wide range of exposure to tobacco smoke constituents. Smokers were enrolled for 19 days and had two periods of confinement in the clinic. Whilst in the clinic, all cigarette filters were collected (for filter analysis and cigarette consumption data) along with 24 hour urine, saliva and blood samples for bioanalysis. Non-smokers provided biological samples only, to indicate background levels of biomarkers in the population.
Filter analysis was used to estimate MLE of nicotine, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), acrolein and pyrene and urine samples were analysed for the corresponding biomarkers as shown in the Table 1. Saliva and plasma were analysed for the nicotine metabolite cotinine.
Table 1: Smoke constituent, associated urinary biomarker and correlation between MLE and BoE measures
Smoke constituent (MLE)
|Urinary biomaker||Pearson Correlation coefficient (r)|
|Nicotine||Total nicotine equivalents (Tneq)*||0.83|
For each smoke constituent, urinary biomarker level was plotted against MLE and these scatter plots are shown in Figure 1. Regression analysis showed strong positive linear relationships across all biomarker/MLE correlations with p < 0.001 in all cases. This demonstrated a relationship between MLE and associated biomarkers. The Pearson correlation coefficients are shown in Table 1.
Figure 1 - Urinary biomarkers vs mouth level exposure (per 24 hours).
When the exposure estimates were examined by group (10mg, 4mg, 1mg and non-smokers) both methods showed a clear dose response, that being for all smoke constituents and their biomarkers the levels remained in rank order with ISO tar yield, i.e. 10mg > 4mg > 1mg > non-smokers. Examples of these plots, for nicotine only, are shown in Figure 2. All other smoke constituents and the corresponding biomarkers showed similar trends.
This study concluded that although biomarkers are the generally more accepted approach for estimating human smoke exposure, filter analysis may provide a simple and effective alternative in studies evaluating alternative machine smoking methods, smoke exposure estimation in medium-and large-scale studies and when evaluating potential combustible MRTPs. Similar conclusions were published by Pauly et al. who reviewed cigarette-filter-based assays as proxies for toxicant exposure. In addition, the ability for these methods to separate different exposure levels provides confidence that these are appropriate tools to evaluate novel reduced toxicant prototype products with reduced yields of specific smoke constituents.