Friday, January 14, 2011

How Could a Study Conclude that Secondhand Smoke Causes Increased Blood Pressure in Children Without Controlling for Salt Intake?


And Why Didn't the Authors Conclude that Secondhand Smoke Does Not Cause Asthma in Children?


It is still not clear to me how a study examining the relationship between parental smoking and blood pressure in children could conclude that there is a causal relationship between the two without controlling for (or even considering) salt intake. It is highly likely that children of parents who smoke also have a higher salt intake. And if that is true, then they are also going to have higher blood pressure. Since the magnitude of the increased systolic pressure among children of smokers was very small (only 1 mm Hg), even a small difference in salt intake between the groups could explain the observed finding.

I find it odd that a paper would conclude, based only on a single cross-sectional study, that secondhand smoke exposure increases blood pressure, especially when the study did not measure salt intake.

Furthermore, it is difficult for me to understand why no one recognized the problem with the study's assessment of asthma medication use. 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. Did no one notice this?

Perhaps the most telling oddity with this paper is its failure to conclude that secondhand smoke is not associated with asthma in children. Since the paper found no association between parental smoking and childhood asthma, why would it not also conclude that parental smoking is not associated with childhood asthma? It appears that the paper is being very selective in its conclusions. Essentially, it is cherry-picking the findings that it likes, and ignoring the findings that it doesn't like.

Now I do not believe that this research supports the conclusion that there is no association between secondhand smoke and asthma. I think a more likely explanation is that the study did not properly assess asthma in this population. This would explain the near impossible finding that only 0.07% of the children took asthma medication.

But since the study authors apparently are confident in their assessment of asthma and asthma medication use (so much so that they are willing to conclude that parental smoking causes increased blood pressure), then why aren't they willing to also conclude that parental smoking is not associated with childhood asthma?

The Rest of the Story

The rest of the story is that I detect a significant bias here. It really appears that the researchers are aiming to link secondhand smoke with increased blood pressure, rather than to objectively examine the relationship. They are cherry-picking from their findings and only giving credence to the finding that goes along with their apparently pre-determined conclusion, and not to the finding that goes against what they apparently want to find.

This bias has become increasingly apparent over the past few years and seems particularly problematic with studies on the effects of secondhand smoke.

No comments: