It is hard to determine the causal relationship; I think it goes in both directions: those who are closer to death are more likely to retire and those who retire are more likely to die sooner.