Passive smoking causes health damage

Yobbo suggested in response to an earlier post that there was convincing evidence that passive smoking was not linked to higher incidence of tobacco-related disease. This is an important argument in the context of tobacco regulation since one motivation for bans on smoking in bars and in the workforce is to reduce the health costs of smoking.

Yobbo cited a paper in the prestigious British Medical Journal by Enstrom and Kabat (2003) which, on the basis of a study of 118,094 US adults either married to ‘ever smokers’ or ‘never smokers’, concluded that there was no connection between environmental tobacco smoke and tobacco-related mortality.

A difficulty with this study was that it started in 1960 when smoking was widespread in the workforce so that other sources of passive smoking would impact on non-smokers than having a spouse who smoked. In addition the study ignored quitting behaviour. Both of these factors would tend to understate the impacts of passive smoking on health. You really need a more direct measure of exposure to passive smoking.

There were many other criticisms of the Enstrom/Kabat paper. Indeed the paper generated a vast response of furious indignation partly motivated by the fact that the authors received funding from the tobacco companies. In the view of some research funded by the tobacco companies should not even be published because of fears the companies are attempting to implant 'seeds of doubt' that will distract from the impact of well-justified health warnings. The editors of BMJ did not accept this reasoning and nor do I.

If the Enstrom/Kabat studied is to be refuted it needs to be done using scientific reasoning or on the basis of the types of objections raised above.

The Enstrom/Kabat studied has in fact been seriously challenged by subsequent work. In particular the study was challenged by Whincup et al. (2004) who accounted more satisfactorily for passive smoking impacts by linking measures of passive smoking exposure (serum cotinine concentrations) to coronary heart disease and stroke incidence among 4729 male non-smokers. Cotinine is a metabolite of nicotine that can be used to directly measure passive nicotine exposure. Whincup et al. found substantially elevated heart disease risks among non-smokers – at the highest cotinine levels these risks were elevated 58% although there were no consistent linkages with strokes. Even low levels of cotinine concentration indicated much higher relative heart disease risks of 45%.

The use of cotinine measures to assess the impacts of smoking bans has been pursued by authors such as Adda & Cornaglia (2006) to quantify the effects of smoking bans in public places. They found that bans on public transport and in schools decrease exposure of non smokers but that bans in recreational public places increase exposure by displacing smokers to private places where they contaminate non-smokers, especially children. Others studies (Farkas et al. (1999)) argue that workplace and home smoking bans increase incentives to quit smoking.