To the Editor:
We thank Finn Diderichsen, Ingelise Andersen, Theis Lange, and Naja Hulvej Rod for their comments on our article.1 To the extent that we understand all the points raised, we attempt to address them here. The authors indicate that they would expect the relative effect of smoking on chronic disease not to vary across levels of socioeconomic position. This runs counter to our thinking and the results from selected extant studies. As we outlined in the original article,2 in analyses of large, population-based surveys, where self-reported cigarette smoking is held constant, cotinine, a biomarker of tobacco smoke exposure, has markedly higher levels in the most disadvantaged folk relative to the affluent.3 This could be because people from lower social groups smoke their cigarettes differently, including inhaling more deeply and more frequently, and/or by choosing higher tar-containing brands. We, as others have before us, therefore hypothesized that the risk to health of smoking in disadvantaged individuals may be elevated in comparison to the more affluent – that is, smoking is most deleterious in lower social groups. While there is some support for this hypothesis – as we describe in our article – and more recently, for health behaviors in aggregate,4 this is not a universal observation; hence, our interest in further testing it is in a well-powered study. Although Diderichsen et al. claim that by computing hazard ratios in this context our work is less comparable to other findings, our reckoning is that the majority of articles in fact report this measure of risk. In the interests of being as comprehensive as possible, we also computed absolute measures of risk in our article – age-standardized mortality rates.
Diderichsen and others also raise the issue of biased reporting of cigarette smoking habits, which may be most acute in disadvantaged populations and which may provide a further explanation for any effect modification of the health consequences of smoking across social groups. This is precisely why we used blood and salivary cotinine to corroborate self-reported smoking – to the best of our knowledge, the first study to do so in the context of the present research question. The use of this biomarker also has the value of capturing cigarette dose in smokers, as would the analyses we computed by self-reported numbers of cigarettes that can be found in the appendix of the article.
G. David Batty
a and Carlos de Mestralb
From the aDepartment of Epidemiology and Public Health, University College London, London, United Kingdom and
bInstitute of Social and Preventive Medicine, Lausanne University Hospital, Lausanne, Switzerland.
1. Diderichsen F, Andersen I, Lange T, Re Rod NJ. Testing differential associations between smoking and chronic disease across socioeconomic groups. Epidemiology. 2019;e38e39.
2. de Mestral C, Bell S, Stamatakis E, et al. Testing differential associations between smoking and chronic disease across socioeconomic groups: pooled data from 15 prospective studies. Epidemiology. 2019;30:48–51.
3. Jarvis MJ, Wardle J. Marmot MG, Wilkinson R. Social patterning of individual health behaviours: the case of cigarette smoking. In: Social Determinants of Health. 2009.Oxford: Oxford University Press.
4. Foster HME, Celis-Morales CA, Nicholl BI, et al. The effect of socioeconomic deprivation on the association between an extended measurement of unhealthy lifestyle factors and health outcomes: a prospective analysis of the UK biobank cohort. Lancet Public Health. 2018;3:e576–e585.