Direct Effect of Birth Weight on Childhood Blood Pressure: a Causal Mediation Analysis

Details

Ressource 1Download: BIB_BCF2AF7916A4.P001.pdf (388.54 [Ko])
State: Public
Version: author
Serval ID
serval:BIB_BCF2AF7916A4
Type
Inproceedings: an article in a conference proceedings.
Publication sub-type
Abstract (Abstract): shot summary in a article that contain essentials elements presented during a scientific conference, lecture or from a poster.
Collection
Publications
Institution
Title
Direct Effect of Birth Weight on Childhood Blood Pressure: a Causal Mediation Analysis
Title of the conference
SER 46th Annual Program
Author(s)
Chiolero A., Paradis G., Kaufman J.S.
Address
Boston, United States, June, 18-21, 2013
ISBN
1476-6256
ISSN-L
0002-9262
Publication state
Published
Issued date
2013
Volume
177
Series
American Journal of Epidemiology
Pages
117
Language
english
Abstract
Background: Numerous studies have shown a negative association between birth weight (BW) and blood pressure (BP) later in life. To estimate the direct effect of BW on BP, it is conventional to condition on current weight (CW). However, such conditioning can induce collider stratification bias in the estimate of the direct effect. Objective: To bound the potential bias due to U, an unmeasured common cause of CW and BP, on the estimate of the (controlled) direct effect of BW on BP. Methods: Data from a school based study in Switzerland were used (N = 4,005; 2,010 B/1,995 G; mean age: 12.3 yr [range: 10.1-14.9]). Measured common causes of BW-BP (SES, smoking, body weight, and hypertension status of the mother) and CW-BP (breastfeeding and child's physical activity and diet) were identified with DAGs. Linear regression models were fitted to estimate the association between BW and BP. Sensitivity analyses were conducted to assess the potential effect of U on the association between BW and BP. U was assumed 1) to be a binary variable that affected BP by the same magnitude in low BWand in normal BW children and 2) to have a different prevalence in low BW children and in normal BW children for a given CW. Results: A small negative association was observed between BW and BP [beta: -0.3 mmHg/kg (95% CI: -0.9 to 0.3)]. The association was strengthened upon conditioning for CW [beta: -1.5 mmHg/kg (95% CI: -2.1 to -0.9)]. Upon further conditioning on common causes of BW-BP and CW-BP, the association did not change substantially [beta: -1.4 mmHg/kg (95% CI: -2.0 to -0.8)]. The negative association could be explained by U only if U was strongly associated with BP and if there was a large difference in the prevalence of U between low BWand normal BW children. Conclusion: The observed negative association between BW and BP upon adjustment for CW was not easily explained by an unmeasured common cause of CWand BP.
Web of science
Create date
30/09/2013 10:43
Last modification date
20/08/2019 15:31
Usage data