Monday, May 02, 2011

Study Finds Secondhand Smoke Associated With Higher Blood Pressure in Boys, But Lower Blood Pressure in Girls; Concludes Effect is Real Only for Boys

According to newspapers throughout the world, secondhand smoke has been found to cause higher blood pressure in boys, based on a new study presented yesterday at the annual meeting of the Pediatric Academic Societies in Denver.

For example, the Guardian headline reads: "Passive smoking raises blood pressure in boys, study reveals."

The headline in the Mirror read: "Second-hand smoke can raise boys' blood pressure and cause heart disease."

The press release put out by the researchers themselves is titled: "Living with a smoker may raise blood pressure in boys."

The study examined the cross-sectional relationship between exposure to secondhand smoke (assessed by serum cotinine levels) and blood pressure in children ages 8-17 from the National Health and Nutrition Examination Survey (NHANES) for the years 1999 through 2006. The research controlled for the following potential confounding factors: age, gender, body mass index, sodium intake, and physical activity.

One of the major findings was that systolic blood pressure among boys exposed to secondhand smoke was 1.6 mm Hg higher than among boys not exposed to secondhand smoke in the home.

The lead study author stated: "While the increases in blood pressure observed among boys in our study may not be clinically meaningful for an individual child, they have large implications for populations."

The study concluded that: "The potential long-term cardiovascular risk of higher BP among ETS-exposed boys provides further incentive to expand policies that create smoke-free environments for children and adolescents."

What the authors are therefore concluding is that the increase in systolic blood pressure among tobacco smoke-exposed boys, even though it was small (1.6 mm Hg) is enough to increase long-term rates of hypertension in the population, leading to increased rates of cardiovascular disease.

The Rest of the Story

What the press release headline doesn't tell you is that the very same study found that among girls, secondhand smoke exposure was associated with a significant decrease in systolic blood pressure. Girls living with a smoker had systolic blood pressure that was 1.8 mm Hg lower than girls who did not live with a smoker.

Thus, it must also be the case that: "While the decreases in blood pressure observed among girls in our study may not be clinically meaningful for an individual child, they have large implications for populations." The potential long-term cardiovascular risk of higher BP among non-ETS-exposed girls provides incentive to limit policies that create smoke-free environments for children and adolescents."

You see, you can't have it both ways. If secondhand smoke exposure increases systolic blood pressure in boys and you conclude that this small difference in blood pressure is going to affect long term cardiovascular disease risk, then you must also acknowledge that the higher systolic blood pressure among non-exposed girls is also going to affect long term cardiovascular disease risk.

In other words, if the children in this study could be followed for many years, according to the scientific reasoning of the study authors, the cardiovascular disease risks of boys who lived with a smoker would be increased, but the cardiovascular disease risks of girls who lived with a smoker would be decreased, if we accept the findings of this study as accurate and valid.

Since the authors argue that this study provides evidence to support smoking bans, they clearly are concluding that the observed increase in blood pressure among boys living with smokers is a real effect, but that the observed decrease in blood pressure among girls living with a smoker is an invalid finding.

In other words, the authors are concluding that their data and findings with regards to the males in their study are valid, but they are completely dismissing the findings with regards to the females in their study.

In addition, the study conclusion disregards the actual finding of the study regarding the observed effect among females. The study concludes: "ETS exposure was associated with significantly higher SBP [systolic blood pressure] in boys, but not girls."

But that is a false statement of the findings. The results did not merely show that ETS exposure was not associated with higher SBP in girls. Instead, the results showed that ETS exposure was significantly associated with decreased SBP in girls.

The study authors apparently try to explain this away by arguing that somehow, girls are not subject to the same cardiovascular effects of secondhand smoke exposure that boys are. This is hardly a plausible explanation. Instead, it seems like a post-hoc excuse that was concocted in order to allow the investigators to draw the conclusion that they wanted to draw.

Moreover, this would only explain the failure to find an association between secondhand smoke exposure and blood pressure in girls. It would not explain the finding of decreased blood pressure among secondhand smoke-exposed girls. Thus, this explanation is invalid on its face.

I have to say that this appears to be one of the most blatant examples of investigator bias in a research study that I've seen. You set out to examine the relationship between secondhand smoke exposure and blood pressure. You find that exposed boys have higher blood pressure and exposed girls have lower blood pressure. You conclude that your findings for males are accurate and valid and that your findings for females are inaccurate and invalid.

If this doesn't appear to be manipulation of data interpretation in order to reach pre-determined conclusions, than I don't know what does.

The rest of the story is that these study conclusions are strikingly biased. The authors appear to have reached a pre-determined conclusion and they appear to have awkwardly and unconvincingly manipulated the interpretation of the study findings in order to support that pre-determined conclusion.

And to add insult to injury, the conclusions of the study are disseminated widely to the media via the "research by press release" approach that I despise. This work hasn't even been peer reviewed yet, but its conclusions are now being read throughout the world. Most of the readers of these publications will never know the rest of the story.

Scientific research and data interpretation needs to be an objective, systematic, and consistent process. You can't simply accept half of your findings and dismiss the other half, especially without a definitive reason. While this might provide some short-term support for smoking bans, in the long run, this harms the scientific credibility and integrity of the tobacco control movement.

No comments: