The concept of "metabolic harmony" is introduced and conceptualized as the state in which indices of metabolic activity (i.e., serum glucose, cholesterol, systolic and diastolic blood pressure, body mass index) within an individual attain their expected values given the individual values on related variables. Its complement, "metabolic disharmony" (i.e., the extent to which an individual's "profile" of metabolic variables is jointly unusual in relation to their expected values) is operationalized via Mahalanbis' D2 statistic calculated on these indices of metabolic activity (plus age and sex). Analysis of a large (N = 5209) longitudinal (32 years) cohort study shows that, independent of the linear and quadratic effects of the aforementioned metabolic variables, the disharmony index (DI) significantly and strongly predicted hazard of death (χ2 (1) = 20.05, P < 0.00005). That is, each 10 percentile increase in DI was associated with a 6.9% increase in the hazard rate. The association of DI to hazard rate was not materially altered when potential confounders (e.g., smoking status) were added to the model or when all subjects were included by imputing missing data. These results demonstrate that metabolic disharmony is associated with, and may cause, an increased hazard of death. © 2001 Harcourt Publishers Ltd.