Control groups in paediatric epilepsy research: do first-degree cousins show familial effects?