Parental smoking cessation and children's daily smoking: public health implications? commentary on Bricker et al.

Abstract

Bricker et al. (2003) have provided us with an interesting and compelling example of epidemiology applied to the issue of smoking in children. Strong points of the study include its large sample size, the very high retention rate and the longitudinal nature of its data collection. They note correctly that while many studies have documented the important role played by current parental smoking in their children's own smoking behavior, relatively few have examined the relationship of parental smoking cessation to children's smoking. Of those that have, there appear to be methodological issues that limit the strength of previous findings with the most serious problem being the use of children's report as a proxy measure for parental smoking. The overall study design provides a strong platform to determine, for the first time, the prospective relationship between parental smoking cessation assessed prior to the typical age at which children start to smoke, i.e. 8–9 years. In this study, parental smoking data were collected at baseline via a mailed survey when their children were in the 3rd grade with an aggressive follow-up (82.9% of parents eventually responded). The children's smoking behavior was assessed 9 years later when they were in the 12th grade. Significantly, the analysis of cotinine in saliva collected at the time of assessment indicated that misreporting of tobacco use was held to an absolute minimum, 1.2–1.5%. Perhaps the only flaw in the approach taken was that parental smoking behavior was not confirmed biochemically; the social desirability for smoking parents to respond as non-smokers may have been especially high in this study. The intriguing findings from this study were that while the odds of children's daily smoking behavior were reduced in the presence of parental smoking cessation (by 25% if one parent quit and by 39% if both parents quit) compared to the effect of continued parental smoking, children were still at an elevated risk for smoking compared to the risk incurred when both parents were never smokers. Moreover, there did not seem to be any effect for when the parental cessation occurred, before or after birth, thereby leading the authors to speculate that a ‘common genetic susceptibility to become a smoker’ may have been passed from parent to child as an explanation of the residual excess risk. This speculation has support from a previous study of ours in which we presented evidence for a major segregating genetic factor for smoking behavior among families from the Western Collaborative Group Study (Cheng, Swan & Carmelli 2000). There is also a plethora of evidence from twin studies that support the hypothesis of genetic influence on smoking initiation (Li et al. 2003). Evidence from molecular genetic studies appear to have identified several candidate gene variants that are associated with smoking behavior, although many of these remain to be confirmed in independent samples using methods to minimize the impact of population admixture on the observed associations. Another pathway to explain the excess risk for smoking in children of former smoking parents could lie in the environmental context in which families with former smoking parents exist. For example, while parents may have stopped smoking, smoking members of the extended family (aunts, uncles, grandparents) and the social network (neighbors, friends, co-workers) may have continued to smoke. Another possibility is that certain characteristics such as parental problem solving, academic achievement, stress/anger management, other substance use and/or psychopathology may have remained unchanged, thereby still leaving the children of former smoking parents vulnerable to become smokers as older children. The authors have also suggested in their paper that the finding that parental cessation influences the likelihood of their children becoming smokers raises public health implications: ‘ . . . if all smoking parents were to quit by the time their children reached age eight then every year 136 000 youth in the United States would be prevented from becoming daily smokers’. While the basis for this estimation is not made explicit and it ignores the influence of many other causal factors, there is an element of truth to it that raises other, potentially more difficult, issues. For example, what is the psychological impact of knowledge that parental smoking, even prior to birth, increased one's susceptibility to become a smoker? Could members of smoking couples who do not quit early in a child's life be held solely responsible for their child's later smoking? Would children with former smoking parents blame them for their own smoking and not take responsibility for their own smoking cessation? The convergence of public health imperatives, as suggested by Bricker et al., with individual rights and future responsibilities as a parent may warrant careful consideration of the potential implications (Parker 1995).