Thursday, August 04, 2005

New Study Links Secondhand Smoke Exposure with Abdominal Obesity, High Triglycerides, and Low HDL Cholesterol in Adolescents

An article published in this week's online issue of Circulation has reported that secondhand smoke as well as active smoking are associated with metabolic syndrome in adolescents -- which in this study, primarily consisted of central (abdominal) obesity, high triglyceride levels, and low HDL cholesterol levels (see: Weitzman M, Cook S, Auinger P, et al. Tobacco smoke exposure is associated with the metabolic syndrome in adolescents. Circulation 2005; 112:862-869).

The authors found a dose-response relationship between cotinine-confirmed levels of tobacco smoke exposure and presence of metabolic syndrome, characterized by 3 or more of the following: central obesity, high blood pressure, high fasting blood sugar, low HDL, and high triglycerides. Compared to adolescents without tobacco smoke exposure, those exposed to secondhand smoke were 4.1 times more likely to have metabolic syndrome and those who actively smoked were 6.1 times more likely to have metabolic syndrome.

The authors conclude that "this is the first study to demonstrate a dose-responsive, cotinine-confirmed relationship between tobacco smoke and the metabolic syndrome and the first that we are aware of to demonstrate any association between tobacco smoke and the metabolic syndrome in adolescents."

The media widely reported the results of this research as implying that secondhand smoke was a cause of metabolic syndrome in adolescents, and by virtue of this relationship, placed teens at higher risk of later developing diabetes and cardiovascular disease.

The Rest of the Story

Based on my review of this research, there are really two problems:

First, the study did not control for diet and levels of physical activity among these adolescents, which I think could reasonably be expected to be associated both with the likelihood of tobacco smoke exposure and with abdominal obesity, high triglycerides, and low HDL cholesterol levels.

A key factor to recognize is that exposure to tobacco smoke in this study implies a much higher likelihood of having a smoking parent (in the dose-response analyses, it actually implied an absolute likelihood of a smoking parent). And think of the whole constellation of factors that are likely to be more prevalent in families when a parent smokes. One might also expect that overall health concerns are lower in such a family, and the children in such families would be more likely to not get exercise and not eat a "healthy" diet (what I mean here is that they are more likely to have a substantially higher fat intake in their diets).

In other words, one would expect that adolescents who are exposed to tobacco smoke would also have a more sedentary lifestyle with less physical activity and a poorer diet, with higher fat intake.

It would not be unexpected to find that kids who get less physical activity and who have higher fat intake would also be more likely to have fat bellies, high fat levels (i.e., triglycerides), and low HDL cholesterol (which is related to the degree of physical activity). It would not be unexpected, then, to also find that secondhand smoke exposure is related to fat bellies, high fat levels, and low HDL levels.

In epidemiology, this is called confounding, and I view it as a critical concern that precludes one from being able to make a causal conclusion in this study. In other words, while secondhand smoke was clearly associated with metabolic syndrome in this study, I don't think one can conclude that the metablic syndrome was caused by the secondhand smoke. In fact, it may likely have been related to a completely different pattern of behavior and lifestyle, marked notably by decreased levels of physical activity and increased fat intake.

A second problem, which would not be a problem if data had been available on physical activity and diet, is that social class was not accounted for in the study. It appears to me that what is really going on here is largely a class effect. Tobacco smoke exposure is characterizing a class of young people who are also characterized by different behavior and lifestyle patterns - including differences in diet and physical activity.

Some of the effects of these potential confounding variables could possibly be accounted for by controlling for parental smoking in the analyses - but it does not appear that was done.

But perhaps the greatest problem in drawing a causal inference from this paper is that the data do not support the underlying hypothesis regarding a mechanism by which secondhand smoke exposure would cause metabolic syndrome. The paper posits that it is largely through insulin resistance that the effects of secondhand smoke on metabolic syndrome would occur: "it is well accepted that insulin resistance is an underlying force driving the metablic syndrome and its components."

However, the paper actually found no evidence for insulin resistance associated with secondhand smoke or active smoke exposure. While 11.6% of adolescents had high fasting glucose levels, 14.1% of secondhand smoke exposed subjects did, and only 9.6% of active smoking adolescents did. These differences were not significantly different, and there was no trend evident.

Thus, the paper does not provide support for the very hypothesis that would need to be relied upon to draw a causal inference.

Another inconsistent finding in the study was that the prevalence of high blood pressure was actually significantly lower in tobacco smoke-exposed adolescents. However, this finding does not, in and of itself, render a causal conclusion invalid.

It is important to note that I am not faulting the authors of the paper at all. They were careful to point out each of the potential threats to the validity of the study, including the possible confounding effects of physical activity and diet. And they seemed to have been quite careful in not drawing causal inferences in the paper itself. It seems to be the media that played the study up to infer a definitive causal relationship between secondhand smoke and metabolic syndrome. I'm not at all blaming the authors for this. But the end result is that I think the media coverage was potentially quite misleading.

But I do have to point out that while the authors were quite accurate and careful in their presentation of the study results to the media, at least one anti-smoking organization was not. The Campaign for Tobacco-Free Kids jumped on the study, and despite what I feel is an inability to draw any causal conclusions from the paper, immediately issued its own definitive causal conclusion:

"The study 'adds to the body of evidence demonstrating that secondhand smoke exposure is one of the most serious causes of disease in the United States,' said Matthew L. Myers, president of the Campaign for Tobacco-Free Kids. 'It is particularly disturbing because it demonstrates that exposure to smoke as a child could well have long-term heart disease consequences.' "

In my opinion (as a researcher who certainly has not downplayed what I view as the health hazards of secondhand smoke), I don't think the study demonstrates anything of the sort. I have seen no evidence (previously or in this study) that exposure to smoke as a child has long-term heart disease consequences, and I find it irresponsible for an anti-smoking organization to be spreading that message to the public.

The rest of the story suggests that the media, and in at least one case, an anti-smoking group, have been far too hasty in jumping to a conclusion that secondhand smoke causes fat tummies, when in fact, this study could simply be another reminder that if you sit around, don't get exercise, and eat a lot of high-fat food, and especially if you're already overweight or at risk of overweight, then you're at higher risk of developing a constellation of abnormalities that may end up putting you at greater risk of heart disease.

While I would argue that we should certainly be taking aggressive steps to protect kids from secondhand smoke exposure, I don't see this study as adding to the evidence of the health risks of secondhand smoke to children. If anything, the immediate action that should come from this is to try to get kids off of the sofa, away from the TV, out into the parks and playgrounds, and on a more reasonably moderate fat diet.

1 comment:

Cantiloper said...

Excellent analysis of this study! I would differ in one regard however.

You are careful to note that you do not fault the authors of the study, but rather the media coverage of the study and the way organized Antismoking groups like CTFK jumped on it to promote their own views and interests.

I agree the authors were not greatly at fault, but I have come to believe that such researchers do indeed have two very great responsibilities when they conduct such research.

1) They should be VERY clear about their funding and any potential conflict such funding might entail. People with the vaguest tobacco industry affiliations are raked over the coals if they don't fly a flag about them... researchers whose work has depended on or will depend in the future on grants from organizations/foundations/companies dedicated to reducing smoking or promoting smoking bans should note publicly how dependent they are on such funds. Such dependencies can easily lead to either the conscious or unconscious biasing of a study design so that it will produce the "proper" results that will be pleasing to potential or future funders. This is a general comment... not pointed at all at these particular authors.

2) When research deals with possible health effects of secondary smoke and finds results that are inconclusive or have a strong possibility of confounding, the authors should make unusually SPECIAL and STRONG note of these limitations in the abstract and conclusion since these are what many in the media focus on and often the only parts of a study that the general non-subscribed public is able to see verbatim. I would go so far as to say that such studies should include a specific "Statement to the Media" outlining exactly what they did and did not find. Authors of studies in this contentious area should certainly know by now how the media will be interpreting their results: they have a responsibility to do whatever they can to avoid misinterpretations that mislead and frighten people... even if such efforts may jeopardize their future special interest funding.

Michael J. McFadden
Author of Dissecting Antismokers' Brains
http://cantiloper.tripod.com