B633 - GSTM1 and cognitive functioning - 31/03/2008

B number: 
Principal applicant name: 
Prof George Davey Smith (University of Bristol, UK)
Dr Beate Glaser (University of Bristol, UK)
Title of project: 
GSTM1 and cognitive functioning
Proposal summary: 

Concept Specific measure Person Source Time point(s)

GSTM1 Del Child DataBase -

GSTM1 Del Mother DataBase -

Cognitive and neurodevelopmental phenotypes in children, measured on the scales of general cognitive, verbal, perceptual-performance, quantitative, memory and motor development, include the MacArtur Communicative Development Inventory (15 mns), the Denver Developmental Screening test (18 mns), the Wechsler Objective Reading Dimension test (WORD, 91 mns), the Wechsler intelligence Scale for Children (WISC, 8.7years), the Wechsler Objective Language Dimension (WOLD test 8.5 years) and the Counting span test (10 years). In addition, child pervasive developmental (DAWBA, Development and Well-Being Assessment, 91 mns) and autism spectrum disorders will be investigated. Genotypic effects of GSTM1 on cognitive and neurodevelopmental outcomes in children will be studied conditional on maternal smoking before, during and after pregnancy, partner smoking, maternal alcohol consumption during pregnancy and type/duration of breastfeeding. Analysis of cognitive functioning in children will be controlled for by birth weight, gestational age at birth, social communications problems (Social Communications Disorder Checklist) and height. Environmental and demographic factors including socio-economic status and maternal factors will be also included in the analyses as covariates.

Multiple regression models, with appropriate adjustment for covariates will be constructed. In addition to categorical coding (2 df test), genotypic effects will be analysed under recessive, dominant and additive disease model assumptions. Recessive coding corresponds to a GSTM1 present vs GSTM1 null analysis. For categorically coded outcomes also deletion-specific analysis will be preformed. For all nominally significant associations empirical p-values will be derived using permutations to account for increased type I error due to multiple testing.

Date proposal received: 
Monday, 31 March, 2008
Date proposal approved: 
Monday, 31 March, 2008
Primary keyword: