Researchers and laypersons have long argued that stress is bad for health, particularly when responses are large, prolonged, and frequent. By extension, individuals who have the largest and the most prolonged responses are assumed to have worse outcomes than do less reactive persons. Research in animals has been supportive of the connection between stress and poor health, but evidence in humans has been slow to accumulate. The current issue of Psychosomatic Medicine presents a meta-analysis of 33 studies of delayed recovery from stress and its association with poor cardiovascular disease outcomes and all-cause mortality. The analysis supports the contention that slower recovery to baseline after exercise or psychological stress may predict earlier death due to all causes. This finding raises questions for psychosomatic theories of disease and points the direction for further study of how or whether to incorporate reactivity measures into standard risk profiles.