OBJECTIVES: The aim of this study was to investigate possible statistical mediators in a randomized controlled trial of mindfulness-based cognitive therapy (MBCT) on pain intensity in women treated for primary breast cancer. MATERIALS AND METHODS: The sample consisted of 129 women treated for breast cancer, presenting with persistent pain, who were randomly assigned to MBCT or a wait-list control. We previously reported a statistically significant and robust effect of MBCT on pain intensity (11-point numeric rating scale), which was included as the primary outcome. The proposed mediators were mindfulness (the Five Facet Mindfulness Questionnaire), self-compassion (the Short-Form Self-Compassion Scale), and pain catastrophizing (the Pain Catastrophizing Scale). Measurement points included baseline (T1), postintervention (T2), and 3- (T3) and 6-month (T4) follow-ups. All indirect effects of the mediators were tested in separate Multilevel Models, using the product-of-coefficients approach with bias-corrected confidence intervals (95% BSCI). The statistically significant mediators were then included in a multiple mediator model. RESULTS: Statistically significant indirect effects were found for mindfulness nonreactivity (B=-0.17, BSCI [-0.32 to -0.04]) and pain catastrophizing (B=-0.76, BSCI [-1.25 to -0.47]). No statistically significant indirect effect was found for self-compassion (B=-0.09, BSCI [-0.30 to 0.04]). In a multiple mediator model, including mindfulness nonreactivity and pain catastrophizing, only pain catastrophizing remained statistically significant (B=-0.72, BSCI [-1.19 to -0.33]), explaining 78% of the effect. DISCUSSION: The results of the present study may have clinical implications. An increased focus on the proposed mediators may optimize the clinical use of MBCT for persistent pain in women treated for breast cancer.