Monday, January 05, 2009

New Study Warns of Dangers of "Thirdhand" Tobacco Smoke

A study published online ahead of print in the journal Pediatrics warns of the dangers, especially to children, of "thirdhand" tobacco smoke -- residual smoke which can cause exposure long after a smoker has stopped smoking. The study describes thirdhand smoke as "persistently high levels of tobacco toxins well beyond the period of active smoking. These toxins take the form of particulate matter deposited in a layer onto every surface within the home; in loose household dust; and as volatile toxic compounds that “off gas” into the air over days, weeks, and months. Smoking indoors on 1 day thus exposes people to tobacco toxins within that space in the future. We use the new term “thirdhand” smoke to name this complex phenomenon and define it as residual tobacco smoke contamination that remains after the cigarette is extinguished" (see: Winickoff JP, et al. Beliefs about the health effects of "thirdhand" smoke and home smoking bans. Pediatrics 2009; 123;e74-e79).

Regarding the risks posed by thirdhand smoke, the study concludes: "The toxicity of low levels of tobacco smoke constituents has been proved. ... Thirdhand smoke may remain inside even when smoking took place earlier. Similar to low levels of lead exposure, low levels of tobacco smoke markers have been associated with cognitive deficits among children. The highest tobacco exposure levels were associated with the lowest reading scores; however, the lowest levels of exposure were associated with the steepest slope in the decrement in reading levels. These facts underscore the possibility that compounds in tobacco smoke are neurotoxic at extremely low levels...".

The study also concludes that messages about the harms of thirdhand smoke, rather than those of secondhand smoke, should be used to help encourage smokers not to smoke at all in the home or to quit smoking altogether: "Thirdhand smoke health education campaigns might be more powerful motivators for these families than simply reiterating information about visible SHS [secondhand smoke] exposure that most families already know."

This message is already being used in an ad campaign to encourage parents to quit smoking altogether, warning them that there are serious risks of simply having clothing in the house that they wore when they were smoking, even it that smoking occurred outside of the home. One ad shows a picture of a coat hanging on a door hook and warns: "The toxins from cigarette smoke can hurt your children long after the cigarette is out."

The Rest of the Story


While I agree completely that parents should avoid smoking in the presence of their children and ideally, they should try to quit altogether, I am not convinced that the existing scientific evidence supports the conclusion that thirdhand smoke poses a significant public health hazard, even to children living in homes with a smoker.

While offgassing of toxic smoke constituents from surfaces onto which the smoke has settled has clearly been shown to occur, the resulting levels of exposure to these constituents in "thirdhand" smoke is extremely low. And there is no convincing scientific evidence that exposures of this magnitude produce any significant health harm, with the one possible exception being children who have asthma and are sensitive to tobacco smoke.

The Pediatrics article cites just a single study to support its contention that low levels of tobacco smoke exposure are associated with health harm: a study which purportedly showed that tobacco smoke exposure characterized by very low levels of cotinine causes children to have reduced cognitive function, as measured by math, reading, and block design scores (see: Yolton K et al. Exposure to environmental tobacco smoke and cognitive abilities among U.S. children and adolescents. Environmental Health Perspectives 2005; 113(1):98-103).

The rest of the story, however, is that this is a cross-sectional study which failed to consider two plausible alternative explanations why children with higher cotinine levels might have lower cognitive testing scores:

(1) These children are more likely to have been exposed to pre-natal tobacco smoke, which may have caused the observed impairment; and

(2) The parents of these children are more likely to have lower levels of cognitive testing scores and their children's scores are reflecting the parents' cognitive abilities and thus an inherited phenomenon.

Let's deal with alternative hypothesis #2 first. The study accounted for parental education, but it failed to account for parents' cognitive testing scores, which were not measured. Thus, the study cannot rule out the very plausible alernative explanation: the study is merely reflecting inherited similarities in cognitive function between parents and their children.

The study itself dealt with alternative hypothesis #1. For a subset of the sample, pre-natal exposure to tobacco smoke was measured and controlled for in the analysis (see Table 4). Interestingly, when pre-natal exposure was accounted for, the relationship between post-natal ETS exposure and math and block design scores disappeared, and the relationship between ETS exposure and reading scores was substantially reduced (by 30%).

Curiously, the study appears to hide or at least obscure the finding that the relationship between cotinine levels and cognitive function - as measured by math and block design scores - disappeared after controlling for pre-natal exposure and that the relationship between cotinine levels and reading scores was greatly reduced.

Part of the reason why the relationship between cotinine levels and reading scores may not have completely disappeared is because pre-natal exposure was based on parental self-report and was likely to be greatly under-reported. The authors acknowledge this, although it doesn't seem to alter their stated conclusions: "These prenatal data were obtained by parental report and may result in an underestimate of the intensity of tobacco smoke exposure. To confirm the causal role of ETS in diminished cognitive abilities among children, prospective birth cohort studies will be necessary."

The authors of the study acknowledge that because of its cross-sectional nature and these other issues, this study is not sufficient to draw a causal conclusion. They also acknowledge that there is no known biologic mechanism by which post-natal ETS exposure could lead to neurologic damage. Nevertheless, this acknowledgment does not appear to stop the authors from drawing such a causal conclusion anyway (an interesting statement about the bias which is now so apparent in tobacco control research).

The acknowledgment also does not stop the authors of the Pediatrics article from concluding that the effects of low levels of post-natal tobacco smoke exposure on children's cognitive function have been well-established (an interesting statement in itself about the bias of tobacco control researchers in reviewing the existing literature, not only in interpreting the results of their own studies).

Three other points deserve mention.

First, it's not clear to me why thirdhand smoke would be of any significant concern in the first place, since such exposure indicates that parents are smoking in the home and that children are therefore being exposed directly to secondhand smoke. Any health effects from the very low levels of exposure to thirdhand smoke will be dwarfed by the health risks from the substantial exposure to secondhand smoke, so why would we be concerned about the offgassing that is occurring?

It's kind of like telling a smoker to avoid exposure to secondhand smoke. In light of the active smoking, why worry about any small incremental effect of secondhand smoke? What the smoker needs to do is quit smoking if she wants to protect her health. Avoiding secondhand smoke seems to be almost a trivial concern in that situation. So why the emphasis on thirdhand smoke when we know these children are being exposed significantly to secondhand smoke because their parents are smoking in the home?

Second, if one were to counter my above point by arguing that smokers may not be smoking in the home, but may be exposing their children by virtue of residue left on clothing, then I would challenge the advocate of that position to produce evidence that such minimal exposure poses any significant health hazard. Based on my review of the literature, it is merely a theoretical concern.

Third, sending a message that smokers need to be worried about the effects of thirdhand smoke could well undermine efforts to encourage smokers not to smoke in their homes. If smokers are led to believe that even if they smoke outside the home, they will still be exposing their children to toxins, then many smokers may simply decide it is not worth the extra effort to smoke outside the home.

This all-or-nothing philosophy may undermine the harm reduction appraoch, and may actually result in smokers deciding not to bother to smoke outside the home. If their children are going to be exposed to toxins anyway, then why bother going through all that trouble?

Messages like the one in the ad campaign highlighted above could well backfire. Or at least, they might prove counterproductive for some smoking parents.

But most importantly, I question the accuracy and the scientific support behind such a health claim. It's not clear to me that if a smoker is careful never to smoke inside the home, that merely hanging up their coat on a hook inside the home will cause harm to his children from toxins that offgas from the coat. I just am not aware of any evidence that this represents a significant health risk.

Those who know my position understand that I believe the effects of secondhand smoke are enough. I do not believe that we need to invoke ever-so-shaky scientific evidence to try to now scare people about the effects of thirdhand smoke. Not only is this approach scientifically unsupported, but it may also backfire by undermining people's appreciation of the documented hazards of secondhand smoke.

No comments: