Circulating antigen carbohydrate 125 (CA125) has emerged as a proxy of fluid overload in heart failure. This study aimed to evaluate the effect of dapagliflozin on short-term CA125 levels in patients with stable heart failure with reduced ejection fraction (HFrEF) and whether these changes mediated the effects on peak oxygen consumption (peakVO). This study is a post-hoc sub-analysis of a randomized, double-blinded clinical trial in which 90 stable patients with HFrEF were randomly assigned to receive either dapagliflozin or placebo to evaluate change in peakVO (NCT04197635). We used linear mixed regression analysis to compare changes in the natural logarithm of CA125 (logCA125) and percent changes from baseline (Δ%CA125). We used the "rwrmed" package to perform mediation analyses. CA125 was available in 87 patients (96.7%). LogCA125 significantly decreased in patients on treatment with dapagliflozin [1-month: Δ - 0.18, (CI 95% = - 0.33 to - 0.22) and 3-month: Δ - 0.23, (CI 95% = - 0.38 to - 0.07); omnibus p-value = 0.012]. Δ%CA125 decreased by 18.4% and 31.4% at 1 and 3-month, respectively (omnibus p-value = 0.026). Changes in logCA125 mediated the effect on peakVO by 20.4% at 1 month (p < 0.001). We did not find significant changes for natural logarithm of NTproBNP (logNT-proBNP) [1-month: Δ - 0.03, (CI 95% = - 0.23 to 0.17; p = 0.794), and 3-month: Δ 0.73, (CI 95% = - 0.13 to 0.28; p-value 0.489), omnibus p-value = 0.567]. In conclusion, in patients with stable HFrEF, dapagliflozin resulted in a significant reduction in CA125. Dapagliflozin was not associated with short-term changes in natriuretic peptides. These changes mediated the effects on peakVO