From a testing perspective the science (knowledge) of tobacco product characteristics has evolved rapidly over the past 40 years driven, to a large extent, by health concerns and tobacco-related regulations. In the late ‘60s and early ‘70s cigarette analysis was focused on deliveries of ‘tar’ primarily as a result of the 1964 Report of the US Surgeon General. Since that time, the number of smoke chemicals and tobacco constituents that are determined routinely has expanded as a result of Health Canada reporting requirements promulgated in 2000 and the more recent publication of the United States Food and Drug Administration (FDA) list of hazardous or potentially hazardous compounds. Current and potential needs for regulatory reporting have resulted in the development of new methods or the modification of existing methods to improve characteristics, effect efficiencies and/or reduce cost. However, even the most advanced multi-component analytical techniques can not compensate for variability related to sample generation. For example, yields of a number of analytes on the Health Canada reporting list have been demonstrated to be dependent on smoking machine type particularly when cigarettes are smoked under more intensive conditions. This suggests that either smoking machines need to be harmonized over a range of smoking conditions or where differences are known, the type of smoking machine should be specified as part of the analytical methodology.

The analysis of aerosols from new generation products (Heat-Not-Burn and e-cigarettes) presents another set of challenges independent of potential issues related to sample generation. Presently, the tendency has been to utilize standard ISO/CORESTA/Health Canada methods under the assumption that if the method works for a complicated matrix such as tobacco smoke, it should work for less complicated nicotine containing aerosols. This may be true but the assumption needs to be verified. Or, current test methods will have to be validated independently for each ‘new’ nicotine containing matrix.

It is now possible to generate analytical results for potentially thousands of compounds in samples prepared using any number of smoking regimes and at least two smoking platforms. Unfortunately protocols for the evaluation of data from a quality control and/or interpretive standpoint have not kept up with our ability to generate data. So there is the very real danger that we find ourselves in the position of not seeing the forest for the trees. For example, assuming that ‘trees’ are analogous to background noise, does it make sense to push detection limits to the point where ambient conditions are likely to be significant contributors to the analytical result? These are some of the questions that will have to be addressed by the current generation of tobacco scientists.