10 years' follow-up of a cohort of white collar men: are psychosocial risk factors predictive of fatal events?