American Journal of Epidemiology Advance Access originally published online on July 31, 2008
American Journal of Epidemiology 2008 168(6):665-666; doi:10.1093/aje/kwn181
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
LETTERS TO THE EDITOR |
THE AUTHORS REPLY
1 The National Institute of Public Health, University of Southern Denmark, Copenhagen, Denmark
2 The Fertility Clinic, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark
3 Department of Neonatology, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark
(e-mail: henriettesvarrenielsen{at}hotmail.com)
We thank James (1) for his interest in our study (2), in which we found preceding brothers to be associated with a reduced birth weight of subsequent children. We fully recognize that we presented epidemiologic and not immunologic data. However, we trust that using several scientific approaches to look at a possible biologic mechanism strengthens final conclusions. We therefore also agree that direct immunologic data are important to support the suggested maternal immune hypothesis. We are performing such immunologic studies now and look forward to presenting these data soon.
James (1) suggests that maternal testosterone level affects offspring birth weight and offspring sex. If maternal testosterone level tracks over time, this mechanism might create a spurious association between preceding brothers and the birth weight of subsequent children. If testosterone or another unobserved confounder affects birth weight and sex, this will create an association between not only sex of the first born and birth weight of the second born but also birth weight of the first born and sex of the second born because birth weight of the first born and sex of the second born share a common cause in maternal testosterone level. This hypothesis can be tested in our data.
We found no association between birth weight of the first born and sex of the second born. Among 126,978 second-born children, a one standard deviation increase in birth weight of the first born was associated with an odds ratio of 1.01 (95 percent confidence limits: 0.99, 1.02; p = 0.32) for the second born to be a boy. A one standard deviation increase in birth weight z score for sex and week of gestation of the first born was associated with an odds ratio of 1.00 (95 percent confidence limits: 0.99, 1.01; p = 1) for the second born to be a boy. These findings are at odds with the mechanism suggested by James (1). Furthermore, they suggest that our finding—preceding brothers are associated with a reduced birth weight of subsequent children—is not due to unobserved confounding.
 |
ACKNOWLEDGMENTS |
|---|
Conflict of interest: none declared.
|
References |
|---|
 TOP References |
|---|
- James WH. Re: "Brothers and reduction of the birth weight of later-born siblings. (Letter). Am J Epidemiol (2008) 168. 000–00.
- Nielsen HS, Mortensen L, Nygaard U, et al. Brothers and reduction of the birth weight of later-born siblings. Am J Epidemiol (2008) 167:480–4.
[Abstract/Free Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||