IJE Advance Access originally published online on January 24, 2006
International Journal of Epidemiology 2006 35(2):491-492; doi:10.1093/ije/dyi314
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Letter to the Editor |
Reply to P. N. Lee
1 Department of Epidemiology, Michigan State University, College of Human Medicine, East Lansing MI 48824, USA
2 New York City Department of Health and Mental Hygiene, Bureau of STD Control, NY, USA
* Corresponding author. E-mail: breslau{at}epi.msu.edu
Mr Lee1 raises the question of a potential overadjustment in Model 4 of Tables 5 and 6, in which we address questions about the association between maternal smoking during pregnancy and children's IQ.2 Displayed in Tables 5 and 6 of our paper, are four successive regression models, in which we addressed three questions: (i) Is maternal smoking during pregnancy associated with children's IQ? (ii) Does low-birthweight mediate the association? (iii) Is the statistical association highly confounded with other factors that influence children's IQ and thus not causal? Results of Model 4, in which the third question is addressed, show that the association between maternal smoking and children's IQ is nearly obliterated and is no longer significant when maternal IQ and education are added to the equation. Model 4 also includes low birthweight, which, because it occurs after smoking during pregnancy, cannot be a confounder. That inclusion is what motivated Mr Lee's question. We failed to caution that reading Model 4 in isolation from previous models in these tables could indeed lead one to question whether we have overadjusted by including low birthweight in the equation.
A previous model (Model 2) in the tables clearly shows that low birthweight did not alter the association between maternal smoking and children's IQ: the crude association was only negligibly reduced when low birthweight was added to the equation (although low birthweight, an established outcome of fetal exposure to smoking, was associated with IQ deficits).
Because of the importance of the topic, we welcome the opportunity to clarify the results. We present below two tables that address Mr Lee's question. Table 1 shows that the inclusion of low birthweight in Model 4 did not influence the estimates: A model that does not include low birthweight (Model 4a in Table 1) yielded the same results as the model we displayed in the article did (Model 4 in Table 1 which is taken from Table 5 of the article). To illustrate the results more clearly, we present in Table 2 comparisons of mean IQ scores of children whose mothers smoked during pregnancy, who smoked but not during pregnancy and who never smoked, according to level of maternal IQ. The association of maternal smoking during pregnancy and child IQ, observed in the last row of Table 2, nearly disappears, when children are grouped according to maternal IQ levels. What remains is an association between maternal IQ and child IQ.
|
|
References
1 Lee PN. Maternal smoking during pregnancy and offspring IQ. Int J Epidemiol 2006:35:491.
2 Breslau N, Paneth N, Lucia V, Paneth-Pollack R. Maternal smoking during pregnancy and offspring's IQ. Int J Epidemiol 2005;34:1047–53.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  N BRESLAU, N PANETH, V LUCIA, and R PANETH-POLLAK Reply to P. N. Lee Int. J. Epidemiol., April 1, 2006; 35(2): 491 - 492. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||