Thursday, January 13, 2011

New Study Concludes that Secondhand Smoke Increases Blood Pressure in Children; Fails to Adequately Control for Asthma Medications and Salt Intake

A new study published online in the January issue of the journal Circulation concludes that secondhand smoke exposure causes increased blood pressure in children ages 5-6 years (see: Simonetti GD, et al. Determinants of blood pressure in preschool children: The role of parental smoking. Circulation 2011; 1213:292-298).

The study was a cross-sectional examination of 4236 children ages 5-6 years living in southwestern Germany. Researchers correlated the subjects' blood pressure with their exposure to parental smoking, controlling for a number of risk factors, including body mass index, prematurity, low birth weight, parental hypertension, parental education, and maternal smoking during pregnancy.

The results were as follows: "In addition to classic risk factors such as body mass index, prematurity, low birth weight, and parental hypertension, both systolic (+1.0 [95% confidence interval, +0.5 to +1.5] mm Hg; P=0.0001) and diastolic blood pressure (+0.5 [+0.03 to +0.9] mm Hg; P=0.03) were higher in children of smoking parents. Parental smoking independently affected systolic blood pressure (P=0.001) even after correction for other risk factors, such as body mass index, parental hypertension, or birth weight, increasing the likelihood of having a systolic blood pressure in the top 15% of the population by 21% (2% to 44%; P=0.02)."

The study concluded: "In healthy preschool children, parental smoking is an independent risk factor for higher blood pressure, adding to other familial and environmental risk factors."

The study conclusion has been disseminated widely by the media, with headlines such as: "Secondhand Smoke Raises Blood Pressure in Kids" and "Secondhand Smoke Raises Children's Blood Pressure."

The Rest of the Story

There are two potential confounding factors which were either not considered or not adequately considered and which I believe invalidate the study conclusion.

1. The Use of Asthma Medication Which Increases Blood Pressure

One of the two most important potential confounding variables is the use of asthma medication by children. The most common type of medication - beta agonists - are well known to increase blood pressure. It is also well-recognized that asthma is more prevalent among children whose parents smoke. Thus, this is an important confounder.

Children whose parents smoke are more likely to use asthma medications that increase blood pressure. If there were no true relationship between secondhand smoke and blood pressure, one would find an increased blood pressure among children exposed to secondhand smoke simply because they are more likely to use asthma medication, which itself increases blood pressure. This is a classic example of the way in which failure to control for an important confounding variable can lead to the finding of a spurious relationship between an exposure and an outcome.

Although the study considers this potential confounder, it is clear that it does not do so adequately. According to the paper, asthma medication use was ascertained in the study. However, the paper reports that out of 4236 children in the study, only 3 used asthma medication that could increase blood pressure (beta-agonists or corticosteroids). This represents a proportion of 0.07%.

That just 0.07% of 5-6 year-old children in the study population use asthma medications like beta-agonists or corticosteroids is simply not believable. A study in the United States reported that 7.4% of 5-9 year-old children in a relatively healthy population use asthma medications that could increase blood pressure. Thus, it is quite clear that the present study did not adequately assess asthma medication use and did not adequately control for this confounding variable.

Perhaps the most telling finding of the paper is that "No difference in the prevalence of asthma or bronchitis was found for children from smoking or nonsmoking families." This is not believable, given the very strong connection between secondhand smoke exposure and asthma.

If the authors trust their own data, then why did they not conclude that secondhand smoke has no association with asthma? After all, their study findings support the conclusion that there is no connection between parental smoking and asthma in young children. Curiously, they failed to draw such a conclusion in the study, or even to mention it in the discussion section.

I don't think there is any doubt that the paper does not adequately measure asthma or the use of asthma medications and therefore, I think it is clear that the paper does not adequately control for this critical confounding factor.

2. Salt Intake

It is very well recognized that among kindergarten-aged children, salt intake is a very important influence on blood pressure. Based on a 2006 study published in the journal Hypertension, just a 42% difference in salt intake among the exposed and unexposed children could explain the entire observed difference in blood pressure in the study. The study, entitled "Importance of Salt in Determining Blood Pressure in Children," makes it very clear that salt intake has a profound effect on blood pressure in young children.

It is a reasonable hypothesis that children of parents who smoke are more likely to ingest greater quantities of salt in their diets. Since increased salt intake raises blood pressure, salt intake could easily confound the relationship between parental smoking and blood pressure. Once again, failure to consider salt intake could lead to a spurious relationship between secondhand smoke exposure and blood pressure.

This study apparently did not measure or consider salt intake as a potential confounding variable.

The rest of the story is that the conclusion of this paper is premature. While it certainly found an association between parental smoking and increased blood pressure, it is not clear the extent to which that relationship is attributable to the increased use of asthma medications in children of smokers and/or to increased salt intake in these children. Without adequate consideration of these two critical potential confounding variables, the paper cannot credibly conclude that secondhand smoke is the cause of the very small increase in systolic blood pressure observed among the children of smokers in this study.

Unfortunately, the conclusions have already been disseminated through the media and so even if the appropriate analyses are done and the findings are found to be invalid, it will be too late.

No comments: