I was going to mention this, too, since it is potentially one of the bigger factors at play here. Some of the other comments suggest there may be a difference in aging-related lifespan when other factors (e.g. occupational fatality) are corrected for.