Do mutations at the glycophorin A locus i patients treated for childhood Hodgkin's disease predict secondary leukaemia?