Background: The relationship between cardiorespiratory fitness (CRF) and mortality risk has typically been assessed using a single measurement, though some evidence suggests the change in CRF over time influences risk. This evidence is predominantly based on studies using estimated CRF (CRF e). The strength of this relationship using change in directly measured CRF over time in apparently healthy men and women is not well understood. Purpose: To examine the association of change in CRF over time, measured using cardiopulmonary exercise testing (CPX), with all-cause and disease-specific mortality and to compare baseline and subsequent CRF measurements as predictors of all-cause mortality. Methods: Participants included 833 apparently healthy men and women (42.9 ± 10.8 years) who underwent two maximal CPXs, the second CPX being ≥ 1 year following the baseline assessment (mean 8.6 years, range 1.0 to 40.3 years). Participants were followed for up to 17.7 (SD 11.8) years for all-cause-, cardiovascular disease-(CVD), and cancer mortality. Cox-proportional hazard models were performed to determine the association between the change in CRF, computed as visit 1 (CPX1) peak oxygen consumption (VO 2peak [mL•kg-1• min-1 ])visit 2 (CPX2) VO 2peak , and mortality outcomes. A Wald-Chi square test of equality was used to compare the strength of CPX1 to CPX2 VO 2peak in predicting mortality. Results: During follow-up, 172 participants died. Overall, the change in CPX-CRF was inversely related to all-cause, CVD, and cancer mortality (p<0.05). Each 1 mL•kg-1• min-1 increase was associated with a ~ 11, 15, and 16% (all p<0.001) reduction in allcause, CVD, and cancer mortality, respectively. The inverse relationship between CRF and allcause mortality was significant (p<0.05) when men and women were examined independently, after adjusting for years since first CPX, baseline VO 2peak , and age. Further, the Wald Chi-square test of equality found CPX2 VO 2peak to be a significantly stronger predictor of all-cause mortality than CPX1 VO 2peak (p<0.05). Conclusion: The change in CRF over time was inversely related to