Stroke genetics – an overview

Document Sample
Stroke genetics – an overview Powered By Docstoc
					Australian Stroke Genetics Collaborative Group
of the

International Stroke Genetics Consortium
Newcastle Chris Levi, John Attia, Rodney Scott, Lisa Lincz, Pablo Moscato, Mark McEvoy, Patrick McElduff Jon Sturm, Jane Maguire Simon Koblar, Jim Jannes Graeme Hankey, Ross Baker

Gosford Adelaide Perth

NHMRC funding for 2009/10/11
• Grants announced October 16th • $1,108,000 funding awarded to the collaborative • Thanks
– John Attia & the Australian team – Hugh Markus & the ISGC

• Stroke Neurology
– Levi, Hankey, Sturm, Jannes

• Genetic Epidemiology
– Attia

• Molecular Genetics
– Scott, Koblar, Lincz

• Biostatistics and Bioinformatics
– Moscato, McElduff, McEvoy

• Project management
– Maguire

• Newcastle/Gosford
– Candidate gene association studies - NHF
• Thrombotic system - stroke outcome

– GWAS macular degeneration - NHMRC

• Adelaide
– Candidate gene studies
• Lacunar stroke

• Perth
– Candidate gene studies - NHF
• Homocysteine & thrombotic system

Australian GWAS
• Original plan
– Single phase study – Currently N= 1,500 cases, 1,500 controls – ongoing recruitment and expansion to 2,000 cases and controls by May 2009 – Aims
• Identify SNPs associated with stroke occurrence and stroke outcome • Translation to the clinical arena
– define whether the genetic signals identified will provide incremental information beyond that provided by known cardiovascular risk factors

– Hypotheses
• A number of SNPs will show strong and consistent association with stroke occurrence and functional outcome post-stroke • These SNPs identified will be useful for clinical risk stratification by
– Adding to the predictive ability of a model using the traditional clinical risk factors – Showing gene-environmental interaction effects – Creating high risk profiles for stroke occurrence and poor stroke outcome, based on the joint effect of genetic and clinical variables

• • • • First ischaemic stroke – CT/MRI Age, gender, and premorbid living arrangements, premorbid mRS. Ethnicity (self-identified) Previous history or current treatment for co-morbid conditions hypertension, ischaemic heart disease, congestive cardiac failure, atrial fibrillation, peripheral vascular disease, diabetes mellitus, hyperlipidemia, known asymptomatic carotid stenosis, and alcohol and cigarette consumption, dietary fat intake, daily exercise activity level Family history of stroke (first-degree family member) and the age(s) at which stroke events occurred in that group. Medications Body mass index (BMI) Blood pressure Stroke severity and mechanism: NIHSS at baseline TOAST Pathophysiological/Imaging based subtypes Stroke outcome assessment: 3-6 month mRS (+ Bartel index, GOS, AQoL in Newcastle/Gosford cases)

• • • • •


• Newcastle/Gosford
– From the Hunter Community Study cohort, a longitudinal study of randomly sampled community-dwelling healthy residents, aged 55 years and older

• Adelaide
– From the Population Research and Outcomes Studies Unit, South Australian Department of Health using computer-aided telephone interviewing for random selection of households listed in the South Australian electronic white pages

• Perth
– From the Perth Community Stroke Study cohort, a longitudinal study of randomly sampled community-dwelling healthy residents

Aim 1
• Genotyping – 1,500 cases & controls Illumina Hap610 arrays (funding secure)
– Illumina’s proprietary software - first pass assessment – Wacholder’s false positive discovery rate test – Benjamini-Liu and Benjamini-Hochberg tests identify SNPs of interest both individually and when they are in linkage disequilibrium

• Validation of top ranking SNPs (funding not yet secure)

Aim 2
• Predictive modelling - adding genetic data to a model using traditional clinical risk factors.
– standard multivariate logistic regression – ROC analysis and comparing areas under the curve

• Gene-environment interaction effects
– Common genotypes – low order interactions - sufficient power using interactions terms in the logistic regression – Higher order interactions (three ways or higher), or interactions with low prevalence exposures or rare genotypes
• Classification and regression trees • Multi-factorial dimensionality reduction • Bioinformatics approaches

Detectable ORs
Allele Frequency Gene main effect Envir main effect
10% prev 20% prev

Gene-Envir interaction
G+E main effect 1.5, Env 20% prev










N=1,300 cases & controls, power 80%, additive or dominant model


• • Funding release - February 2009 Steering Group establishment – November 2008
– – – Governance structure International representation On-going links with ISGC Research plan refinement Sub-studies Collaborations Site visits Database development Phenotype assessment “quality control” across sites DNA sample quality control


Project Plan review – December 2008
– – –


Project Manager appointment – February 2009
– – – –

• • •

On-going recruitment case and control samples to reach 2,000 by end 2009 Laboratory analysis commences – Mid 2009 Laboratory analysis completed – Mid 2010