It is widely acknowledged that there is a strong need to identify which modifiable risk factors predict healthy aging to use this information as the scientific basis for systematic interventions. Data from a 4-year longitudinal study on aging among 5632 older Italians were used. The definition of vitality was based on both cognitive and physical status, and the envisaged transitions were: positive or nonpositive stable, positive or negative transition, lost, and deceased. Predictors associated with different vitality trajectories were investigated by multinomial logistic analysis with a six-level outcome. Age and educational level were predictors of being "positive stable," whereas the other factors behaved differently according to comparison group. For example, being overweight is a common predictor except when compared to the deceased group, as is depressive symptomatology except when compared to the "positive transition" group. Interventions are warranted to reduce social inequalities, promote adequate body weight, and prevent and treat depressive symptoms.